MINISTRY OF EDUCATION AND TRAINING

MINISTRY OF NATIONAL DEFENCE

ACADEMY OF MILITARY SCIENCE AND TECHNOLOGY

BUI NGOC THUY

Research ON THE ACCURACY IMPROVEMENT OF THE TARGET

PARAMETERS IDENTIFICATION AND DETERMINATION USING

POLARIMETRIC SYNTHETIC APERTURE RADAR AND

POLARIMETRIC INTERFEROMETRIC IMAGES

Specialization: Electronic engineering

Code:

9 52 02 03

SUMMARY OF DOCTORAL THESIS IN ENGINEERING

Hanoi, 2019

This thesis has been completed at:

ACADEMY OF MILITARY SCIENCE AND TECHNOLOGY

Scientific Supervisors:

1. Assoc. Prof, Ph.D Le Vinh Ha

2. Ph.D Pham Minh Nghia

Reviewer 1: Prof, Ph.D Bach Gia Duong

University of Engineering and Technology, Vietnam

National University, Hanoi

Reviewer 2: Assoc. Prof, Ph.D Hoang Van Phuc

Military Technical Academy

Reviewer 3: Ph.D Le Thanh Hai

Academy of Military Science and Technology

This thesis was defended at the Doctoral Evaluating Council at Academy

level held at Academy of Military Science and Technology

in 8:30, date ….month ….. year 2019.

The thesis can be found at:

- Library of Academy of Military Science and Technology

- Vietnam National Library

1

INTRODUCTION

1. The urgency of thesis

Over the last two decades, with the development of science and technology,

remote sensing data has been used more widely in many fields. Remote sensing

applications have been supporting effectively in tactics and modern warfare by

creating the unexpected elements, surveillance and locate targets exactly, to name

but a few. By processing remote sensing data that are obtained from satellites and

aeronautics with very high resolution, the identification and determined of objects

such as military targets and civilian targets could provide the relative accuracy

without making physical contact with the objects.

Particularly, in order to assess the impact of changing forest ecosystems, ice,

as well as the phenomenon of global climate change, remote sensing technology

has been researched and developed and applied in varied fields of socio-economic

life. Technique of Polarimetric Synthetic Aperture Radar and Polarimetric

Interferometric (PolSAR and PolInSAR) is a topic that is drawing the attention of

scientists all over the world. This technology contributes to deal with the problems

of the object‘s identification and measuring target parameters. Therefore, the

thesis has its scientific and practical significance.

Therefore, the PhD student has chosen the topic: Research on the accuracy

improvement of the target parameters identification and determination using

polarimetric synthetic aperture Radar and polarimetric interferometric images.

2. The objective

Research scientific basis and solutions to improve the accuracy of

identification and identify parameters of natural targets, targets and forest height

estimates based on PolSAR and PolInSAR images.

3. The subject and scopes

PolSAR's target decomposition technique is based on Freeman's threecomponent scattering model and Yamaguchi's four-component scattering model.

The PolInSAR decomposition technique is based on the coherence set and the

three-component decompositon technique for PolInSAR images. From above

researchs, proposing and developing solutions and simulations for the accuracy

improvement of interpretation, identification and determination of target

parameters are studied by the thesis.

4. Research methodology

- Methods of collecting information, documents, general analysis of scientific

works and articles published in the world and in the country. Collect PolSAR and

PolInSAR image data sources related to the test area.

- Researching target decomposition techniques and building algorithm models

to improve the accuracy of identification and determination of target parameters

based on PolSAR and PolInSAR images.

- Programming technology and application of informatics technology in

building a program to perform calculations and simulations using MATLAB tool

2

combined with specialized software ENVI 5.0 and PolSARproSim, performing

verification experiments.

5. Scientific and practical significance of the thesis

- The research results of the thesis have contributed to the decomposition

technical theory based on the three-component scattering model and four

components for PolSAR image with asymmetric scattering model to improve the

ability to identify natural objects, artificial targets and to improve the accuracy of

estimating forest height in PolInSAR images.

- The research results of the thesis provide a full assessment of the scientific

basis as well as a test result of solutions to improve the identification accuracy and

identify targets that can be applied for military and civil purposes. Serving

teaching research, specialized research, realizing applied software and developing

remote sensing technology in Vietnam.

6. Structure of the thesis

The thesis is composed of the beginning, 3 chapters, the conclusion is as

follows:

Chapter 1: Overview of identification and determination of target’s

parameters based on polarized radar and polarization-interference images; Chapter

2: Propose target identification algorithm based on PolSAR image; Chapter 3:

Estimating target parameters based on PolInSAR images; Finally, the conclusions,

assessments and the issues that need further research.

Chapter 1. OVERVIEW OF IDENTIFICATION, DETERMINATION

OF TARGET PARAMETER BASED ON POLARIZED RADAR

AND POLARIZATION INTERFERENCE IMAGES

1.1. Radar equation

The electromagnetic waves of radar system transmitted in the medium can be

obtained at a target and the energy carried by the incident wave is absorbed by the

target itself, whereas the rest is reradiated as a new electromagnetic waves. Due to

the interaction with the target, the properties of the reradiated waves can be

different from those of the incident ones. From this change, they can help us

describe or identify targets. In reality, we are interested in the changes concerning

the polarization of the wave.

Figure 1.1: The interaction of electromagnetic waves with a target

3

The radar equation represents as the following:

P G ( , ) A ( , )

(1.1)

PR T T 2 ER 2

4 rT

4 rR

1.2. The formation of Synthetic Aperture Radar image

SAR image systems allow for monitoring the earth on a global scale at all

times and in all weather conditions. The basic geometry of a SAR system is shown

in Figure 1.3.

Figure 1.3. Basic geometric

Figure 1.4. Ground range to slant

structure of a SAR space

range projection

One of the most important criteria for assessing the quality of SAR image

systems is its spatial resolution. The spatial resolution describes the visibility of

the image as much as possible so that two scattered objects can be separated in

term of spatial meaning.

The imaging SAR system is a side-looking radar sensor with an illumination

perpendicular to the flight line direction as shown in Figure 1.4.

1.3. Characteristics of target polarization.

1.3.1. Polarized information

Polarized information in backscattered waves from a given environment can

be related to: reflecting geometric structures such as shape and orientation or

geophysical structure such as moisture, surface roughness face…

(a) Polarimetric SAR (PolSAR)

(b) Single- polarization SAR

Figure 1.8: Polarizing types in remote sensing radar

4

1.3.2. Target scattering vector k

The relationship between incident and scattered waves is as follows [42]:

ES

e jkr

e jkr S11

S EI

r

r S21

S12

EI

S22

(1.12)

To be able to extract physical information from the 2×2 coherent

backscattering matrix or Sinclair matrix which achieved through the construction

of system vectors, we represent the Sinclair matrix with the vector V(.) as follows:

S

S HH

SVH

S HV

SVV

1

k V S Trace S

2

(1.13)

Where ψ is a set of 2×2 complex basic matrices which are constructed as an

orthogonal set under the Hermitian inner product.

1.3.3. Coherency matrix [T] and covariance matrix [C] polarimetric

The polarimetric Pauli coherency matrix [T] and the Lexicographic

covariance matrix [C] are generated from outer product of the associated target

vector with its conjugate transpose as.

T

k3P .k 3*TP

and C k3L .k 3*TL

(1.34)

Where superscript T and * denote transpose and complex conjugation and

denotes the ensemble average in the data processing, respectively.

1.4. Scattering mechanisms

- Surface scattering.

- Double-bounce scattering.

- Volume scattering.

- Helix scattering.

- Wire scattering.

1.5. Target polarimetric characterization

Symmetric scattering hypotheses about the distribution of scattering objects

will make scattering problems simple and allow for quantitative conclusions about

their scattering properties.

1.6. Polarimetric target decomposition PolSAR

Model-based decompositions for the target based on physical scattering

models to interpret the scattering process.

(1.67)

C CS CD CV

In addition, the total power P of all components can be retrieved by

summation of the power contributions PS, PD, PV, which are calculated from the

trace of the single scattering component matrices.

(1.68)

P Tr CS Tr CD Tr CV PS PD PV

Finally, the normalized power of each scattering component can be obtained

by division with P. This provides the opportunity to compare the strength of the

different scattering contributions with respect to each other.

5

1.7. Interferometry Synthetic Aperture Radar

Interferential Synthetic Aperture Radar (InSAR) takes advantage of the phase

difference between two SAR complex images described from differences at

locations or at different times.

Figure 1.21: Geometric structure of interference radar

When building the complex product s1s2* for interferometry, we could cancel the

scattering phase terms and keep the geometrical phase. In effect, we now obtain a

signal phase which depends only on the difference in range between the two positions

R R1 R 2 as follows:

j 2 2 R1s1

4

j

R

s1 a1e

*

(1.71)

s

s

Ae

1

2

2

j 2 R2 s2

s2 a2 e

The observed component in these interferometric radar systems is the

interference phase and is determined as follows:

4

(1.72)

arg s1s2*

R 2 N ; N 0, 1, 2,...

1.8. Polarimetric Interferometric Synthetic Aperture Radar

PolInSAR differs from conventional SAR interferometry in that it allows

generation of arbitrary interferogram and receive polarization pairs as Figure 1.22.

Using the outer product formed from the scattering vectors k1 and k2 for

images S1 and S2, we can define a 6×6 Hermitian positive semidefinite matrix [T6]

and [C6] as follow:

k

3 P1 k *T

T6

3P1

k

3P2

k3*T

P

k

3L1 k *T

C6

3L1

k

3L2

k3*T

L

T1

2

*T

2

C1

C *T

int

T2

Cint

C2

(1.76)

(1.77)

6

The interference equation is as follow:

arg 1k3P k 3*TP *2T

1

2

arg

*T

1

2

(1.80)

By using Eq. (1.85), the complex interferometric coherence as a function of

the polarization of the two images may be written as:

12*

1*2

1, 2

(1.81)

11* 2 2*

1*T T1 1 *2T T2 2

Figure 1.22: PolInSAR acquisition geometry

1.9. Conclusion of Chapter 1

Through an overview of the target decomposition technique based on PolSAR

and PolInSAR images, the works have been published in foreign and domestic

scientific journals as well as factors affecting the problem of identification and

correctly determination the target parameters shows:

- Research on resolving disadvantages in pixels still has many negative power

components, this is the main cause leading to inaccurate identification of targets.

- Volume scattering components are often assumed to be symmetric reflective

scattering, the number of observations is limited because some assumptions must be

accepted to eliminate ambiguity, assumptions often cause the negative power in the

scattering mechanisms leads to incorrect assessment of forest height.

Therefore, the thesis is to set out 2 main problems to solve:

Building PolSAR image processing algorithms based on scattering models in

response to the requirements of interpretation capacity enhancement, target

parameters determination such as the accuracy of natural and artificial targets

identification. This problem is solved in chapter 2.

A proposed algorithm to improve the accuracy of forest height estimation

based on simulation and satellite data sources. This problem is solved in chapter 3.

7

1

Chapter 2. A PROPOSED ALGORITHM FOR IDENTIFYING

TARGET BASED ON POLSAR IMAGE

2.1 Targeted decomposition techniques based on polarized radar images

2.1.1. Technical analysis of coherence target

The scattering model after rotating an angle can be expressed as simple

conversion as in the following expression (2.1):

sin

cos sin SHH SHV cos

S

(2.1)

cos SVH SVV sin cos

sin

Using a unitary transformation matrix according to the parameters of the Pauli

rotation matrix, we can express them in a form of unitary vector as follows:

1

1 0

k

2 0

0

0

cos 2

sin2

0

0

sin2

cos 2

0

0 S HH SVV

0 S HH SVV

0 S HV SVH

1 S HV SVH

(2.2)

We find that the complex totals SHH SVV and SHV SVH are invariable for the

axis of rotation, which gives them a special physical meaning.

2.1.2. Technical analysis of targets according to the scattering model

The decomposition technique following the basic scattering model as shown

in Figure 2.2 including scattering of the surface, double bounce and volume

scattering:

Figure 2.2: Basic scattering models

2.1.2.1. Freeman-Durden three component decomposition

The Freeman-Durden (FDD) decomposition is a technique for fitting a

physically based, three component scattering mechanism model to the PolSAR

observations, without utilizing any ground truth measurement.

The Freeman-Durden decomposition model:

C3V C3S C3 D C3V

f 2 f 2 3 fV

D

S

8

0

f * f * fV

D

S

8

0

2 fV 8

0

8

0

3 fV

fS fD

8

f S f D

fV

(2.17)

8

Equation (2.17) gives us four equations with five unknowns. However, the

contribution of a number

fV 2 fV

3f

,

or V of volume scattering can be eliminated

8 8

8

2

2

*

, then we get 3 equations in 4 unknown equations:

SHH , SVV and S HH SVV

*

S HH S HH

fS fD

2

2

*

S HH SVV

f S f D

(2.18)

*

SVV SVV

fS fD

The fS, fD coefficients and or parameters can be used to determine the

properties of the target. Finally, the contribution of each scattering mechanism can

be determined for the following spans:

Span SHH 2 SHV SVV PS PD PV

2

2

PS f S 1

With:

2

;

2

PD f D 1

2

;

(2.19)

PV fV

(2.20)

2.1.2.2. Yamaguchi decomposition

The three-component scattering model using a covariance matrix is very

effective for scattering mechanisms in PolSAR and the algorithm is only based on

*

*

symmetric reflections SHH SHV SHV SVV 0 , which is the limitation of the FDD

*

*

technique [12], due to SHH SHV 0, SHV SVV 0 in reality.

The Yamaguchi decomposition model:

C f S CS f DCD fV CV f H CH

2

fS 0

*

0

0

0 d

0

0

b

0

2

0 fD 0

*

1

0

a

0 fV 0

1

d

0

1

0 fH j 2

c

1

j 2

2

2*

1

j 2

(2.28)

1

In which, fS, fD, fV, fH indicate the determined scattering coefficients,

corresponding to the surface scattering, double bounce, volume and helix

scattering mechanisms.

The contribution of each scattering mechanism can be estimated as:

2

2

PS f S 1 ; PD f D 1 ; PV fV a b c ; PH f H

(2.29)

2

2

2

Pt PS PD PV PH SHH 2 SHV SVV

2.2. Target identification based on the three-component scattering

decomposition with an adaptive volume modelling

Adaptive algorithm solves the problem of general eigenvalues analysis and

non-negative power constraints, thereby determining the unique minimum value

for volume scattering. Finally, we determine the power of two remaining

scattering components. The generalized double-bounce scattering component

reflects the interaction of electromagnetic waves in both natural and urban areas.

9

* Proposed decomposition method:

The coherence matrix of target is analyzed into a combination corresponding

to a practical scattering mechanism.

Then [T] by given as

1

PS

T f S TS f DTD fV TV

2

1

0

*

0

2

0

P

0 D

0

1

*

2

*

*

a d 0

P

* V d b 0 (2.40)

2

0 0 c

2

=1+ ; a b c

(2.41)

With , , are parameters of the surface scattering and double bounce

scattering models

When is known, we can calculate the remaining power Pv, and PS.

.P 2

P P

2

PD Pt ; PV (T33 t . ); PS 1 T11 D V a

(2.49)

c

Algorithm flowchart:

2

2

The coherence matrix [T]

Determined TS, TD, TV

0.3

PV 0; PD .Pt

PS 1

2

T

11

*

T23

T33

;

PD

T33

*

;

PD 0; PV c

PS 1

2

T

11

PV

0; 0;

T13

PD 2

P

T11 D

T22

PV

b

P

T11 V a

T22

Calculate PS., PD and PV using parameters , and

Figure 2.5: The flowchart of proposed algorithm

To evaluate the proposed decomposition method, the experiments were

performed using the full PolSAR data of ESAR (Experimental Synthetic Aperture

Radar) with 3x3m resolution. By testing the area near Oberpfaffenhofen,

10

Germany, Figure 2.6 (a) shows the optical image of this area. The is the ratio of

asymmetric scattering power to total power. Figure 2.6 (c) shows the color image

of the coefficient.

Figure 2.6: Survey area, (a) Optical image of Oberpfaffenhofen area,

(b) Pauli image, (c) Image color of correlation coefficient

A

B

The proposed

decomposition

(a)

(b)

Freeman decomposition

Figure 2.7: Decomposition image of the test area, (a) Color image of the three

components of the proposed decomposition method, (b) Color image of the three

components of the Freeman decomposition method

* Experimental results:

The effectiveness of the proposed method is evaluated in comparison with

Freeman's three-component decomposition. From the compared results, we can

evaluate accurately with dominant double bounce scattering mechanisms in the

urban area and dominant volume scattering in the forest area. In the figure 2.7 (a),

the volume scattering component determined from the proposed decomposition

shows that the observation is clearer than that in the Freeman decomposition

technique as shown in Figure 2.7 (b). In Figure 2.7 (b) many pixels in the image of

the urban area are still green, thus lead to misinterpret and misidentify the target.

The main reason is the existence of many pixels with negative power components

in the Freeman decomposition technique.

11

In Figure 2.8, we could find that the value is quite low corresponding to

forest area and agricultural land. However, in areas with artificial structures or

urban areas, the value is quite large.

Figure 2.8: Graph at the test areas (a) forest area,

(b) farmland area, (c) urban area.

Thus, the proposed algorithm has added the asymmetric double bounce

scattering component, consequently, the image results have significantly

improved, and the target identification and determination are more precicely than

that in the Freeman decomposition.

Forest

Agricultural land

Urban

The proposed method

(a)

(d)

Surface scattering

(b)

Freeman decomposition

(e)

Double bounce scattering

(c)

(f)

Volume scattering

Figure 2.9: Pie chart of three components of scattering in surveyed areas.

(a,b,c) the proposed decomposition method, (d,e,f) Freeman decomposition

12

It could be seen that the stability of the proposed decomposition method

compared to the Freeman dicomposition method is more clearly in comparing the

ratio of pixels with non-negative power components to filtering windows of

different sizes, shown in the Table 2.1.

Table 2.1: The table of the ratio of pixels with non-negative power components

Filter window size

1×1

3×3

7×7

9×9

Proposed method

99.09%

99.14%

99.20%

99.32%

Freeman method

97.05%

98.21%

98.64%

98.83%

From the above results, comparisons and analysis, the proposed method is

better and more stable than Freeman's three-component decomposition method.

Although asymmetric double bounce scattering could not completely solve all the

scattering problems in urban areas. Some pixels of buildings and the transport

system are still misinterpreted and misidentification.

2.3. Urban target identification based on four-component decomposition

technique with extended volume scattering model

In the proposed method, the thesis uses adaptive algorithms to determine the

covariance matrix parameters of volume scattering and asymmetric scattering. The

effectiveness of the proposed method is proved with PolSAR data obtained from

ESAR airborne remote sensing system.

* Proposed decomposition method:

To better describe asymmetric scattering, the method will take into account

all parameters related to asymmetric scattering information, the proposed

decomposition model is presented as follows:

C f S CS f DCD fV CV f asymCasym

2

f S 0

*

0

0

0

2

0 f D 0

*

1

0

0

0

a

0 fV e*

*

1

d

e

b

f*

d

f

c

f asym

2

2

*

2 *

*

2

2

2 *

2

1

(2.54)

2

2

where 2 1; a b c; Pasym is the asymmetric scattering

power component; 2 , 2 are both complex numbers corresponding to C12

and C23, The scattering terms can be utilized to describe the general case of

nonreflection symmetric scattering, generally, C12 0 and C23 0 and a f are

the elements in the volume scattering model.

Compared to the Yamaguchi four component decomposition method, the

proposed decomposition method uses the asymmetric scattering mechannism.

Therefore, it can be seen that the asymmetric scattering often appears in complex

urban areas, and disappears in natural distributed areas.

* Experimental results:

The effect of the proposed algorithm is evaluated based on dataset received

from the E-SAR airbonre system and the test area which is close to

Oberpfaffenhofen, Germany.

13

Figure 2.11: Optical image of the survey area (a) Optical image GoogleEarth,

(b) Image color of correlation coefficient

A

B

C

The proposed decomposition

(a)

(b)

Yamaguchi decomposition

Figure 2.12: Image decomposition of the test area (a) The proposed

four-component method, (b) Yamaguchi decomposition

From the comparison of image results shown in Figure 2.12 (a) we could see

that urban areas are purplish red, better than those in Yamaguchi's decomposition

in Figure 2.12 (b). From the results of the proposed decompositon method, the

target identification and determination of the proposed method are more precise

than that in the Yamaguchi decomposition technique.

The proposed method utilise an asymmetric scattering component in order to

improve the limitation of Yamaguchi decomposition in terms of the interpretation,

estimation, and classification of the terrain of the targets. From the above results,

comparisons and analysis, the proposed method shows better improvement of

Yamaguchi's four-component decomposition method. Although the proposed

method still has the disadvantage is the forest areas located bottom right image of

Figure 2.12 (a) misidentification due to asymmetric scattering component exceeds

the power level. In the future, the thesis will continue to research further

improvement and hope for better results aimed at improving the effectiveness of

the proposed method.

14

Forest

Agricultural land

The proposed method

Urban

(a)

(b)

Yamaguchi decomposition

(c)

(d)

(e)

(f)

Surface scattering

Double bounce scattering

Volume scattering

Asymmetric scattering

Figure 2.13: Pie chart of four-components of scattering in surveyed areas.

(a,b,c) the proposed decomposition method, (d,e,f) Yamaguchi decomposition

2.4. Conclusion of Chapter 2

1. The overview of the Freeman and Yamaguchi target decomposition

techniques.

2. Build two target identification models based on three-component and fourcomponent decomposition techniques with asymmetric scattering models.

3. The program is set up in Matlab environment. Then, based on the results

extracted from Matlab software, the thesis continues to use specialized software

ENVI 5.0 to perform interpretation and analysis of target.

4. Experimental results of the two proposed algorithms at the points with a

negative power component significantly reduced. Therefore, there has been a

significant improvement in the identification and interpretation of targets more

accurately than the Freeman decomposition technique and the Yamaguchi

decomposition technique.

From the results of the above study, the proposed methods show improved

accuracy in interpretation, identification and determination of target.

Chapter 3. ESTIMATING PARAMETERS OF TARGET

BASED ON POLINSAR IMAGE

3.1. Methods of height estimating of target

3.1.1. Three-stage inversion technique

Forest height estimation in three-stage inversion technique can be divided

into three separate stages.

15

The implementation of the three-stage inversion method is summarized in

Figure 3.1.

Figure 3.1: Three-stage inversion method

The forest height estimation is usually determined after eliminating the

phase component of the surface in complex interferometric coherence coefficients

of the volume scattering component in terms of an inversion space in the unit

circle. A vol value in the inversion space can be used to recover the pair of value

of height and extintion coefficient.

The definition of average height of tree is:

1 N

hV hV i

(3.2)

N i1

The root mean square error is defined as:

N

RMSE

hV i hV

i 1

N

2

(3.4)

Where hV-i is the tree height estimation at i-th pixel, and N is the number of

samples used for the calculation.

3.1.2. Estimation of Signal Parameters via Rotational Invariance Techniques

The main idea of this method is to transform the observation space into two

interpolated subspace: signal and noise. ESPRIT algorithms can be extended to

analyze SAR data. For SAR data, the backscattering signal received from the antenna

is analyzed into a total of different scattering mechanisms. The interferometry phases

depend strongly on the observed medium.

16

2 Figure 3.2: ESPRIT algorithm

3.1.3. Coherence set theory

The continuous polarization space enables us to use the polarization diversity

and to consider the set of all coherences as a continuous entity. We will refer to

the set of all coherences in the following as the coherence set. It has been

identified that the coherence set is directly connected with the mathematical

concept of the numerical range (also known as field of values) of the contraction

matrix. This section is aim to present some theoretical aspects of the PolInSAR

coherence set. The first part of this section deals with the different definitions of

the numerical range. The second and third parts present principles for information

extraction of the coherence set geometry and the coherent properties in the

complex coherence plane.

3.2. The forest parameter estimation based on the target decomposition

technique for PolInSAR images

In the proposed method, the accuracy of forest parameter estimation can be

improved by a coherence method. First, the phase and the parameters of the

scattering object in the tree canopy are determined by the target analysis algorithm

by the adaptive scattering model, then removing the scattering component from

the covariance matrix and the phase of the terrain is determined by the ESPRIT

algorithm. Finally, the forest height is estimated and compensated by the

coherence compensation method. Experimental results show that the accuracy of

forest parameter estimation is significantly improved.

3.2.1. Polarization interference coherence

The data received from the PolInSAR system are usually represented by a

complex coherence matrix 6×6, and are represented in (3.28).

17

T kk

T1

*T

*T

T2

k

1

k 2

với k

(3.28)

The polarization interference coherence of PolInSAR system could be

described by a polarization function of two images and represented as follows:

1*T 2

*T

(3.29)

1, 2

*T

1*T T1 1 *2T T2 2 T

where 1 2 is a unit complex vector of each polarization channel,

T T1 T2 2 . The contracted form of coherence matrix for PolInSAR data is given by:

T

1 2

T

1 2

11 12

21 22

0

0

0

0

33

(3.30)

3.2.2. Scattering mechanisms for PolInSAR data

- Surface scattering

- Double-bounce scattering

- Volume scattering

3.2.3. Estimating scattering parameters of trees

The contracted coherence matrix PolInSAR's is analyzed into the sum of three

sub-matrices corresponding to the three scattering components: volume scattering,

double bounce scattering and surface scattering:

j

f S e S TS f D e jD TD fV e jV TV

(3.49)

where f S , f D .and fV represent the scattering power coefficient of single bounce,

double bounce and volume scattering respectively.

If 11 < 33 then the parameters of scattering components are determined

as follows:

fV 1

2

11

; V arg 1

2

11

2

2

0 ; f S 0 ; f D 22 33 11 ; D arg 22 33 11

(3.50)

Conversely, if 11 33 , then parameters of random scattering

components from tree canopy are directly determined from (3.49)

fV 33 ; V arg 33

T

TV 33

v33

FS f S e jS

FD f D e

jD

jV

; FV fV e ; arcos

11 22 TV 11 TV 22 FV

11

12

;

1

22 33 FD

22 TV 11 TV 22 FV

2

4 12 TV 12 FV

2

2

11 22 TV 11 TV 22 FV

11 22

2

TV 11 TV 22 FV

2

4 12 TV 12 FV

2

(3.51)

18

3.2.3. Estimate the parameters of terrain using the coherence set

In order to reduce the computational complexity and to improve the

accuracy of the terrain parameter estimation, the thesis proposes the using of the

coherence set to directly estimate the parameters of the terrain.

Algorithm flowchart:

PolSAR

Dataset 1

PolSAR

Dataset 2

Co-register photos

Contraction matrix

Estimating scattering parameters of

trees by 3-component technique

11 33

V arg 1

Estimating surface scattering

parameters using the coherence set

11 33

33

T

V 33

2

11

hV

V arg

V 0

kZ

V 0

0 arg 2 31 L

R sin

4 Bcos

Figure 3.7: The flowchart of proposed algorithm

Based on the characteristics of contracted coherence matrix in (3.31), we have

a coherence set for PolInSAR data as follows:

(3.52)

app w*T w : w*T w 1, w 3

Equation (3.52) has the same form as the numerical range of the square

3

matrix A . Therefore, the numerical range of the matrix can also be

considered as the region of the coherence set.

The interferometric phase of the ground topography is determined as:

0 arg 2 3 1 L

(3.53)

Finally, the forest height is determined by the differency between the

scattering phase of the canopy and the scattering phase of the ground topography,

as shown in (3.56).

0

R sin

(3.56)

hV V

V 0

kZ

4 B cos

19

Where is the angle between the radiation wave and the vertical axis, R

is the distance between radar and target, is the angle of deviation between the

baseline and the horizontal axis, is the electromagnetic wave.

3.2.4. Experimental results

3.2.4.1. Simulation data

The simulated data obtained from PolInSAR system L-band (1.3 GHz)

with 100m horizontal and 10m vertical baseline is shown in Figure 3.8. The test

forest area has an average height of 18m on a relatively flat topography with the

forest stands occupies a 2.6 Ha area and stand density is 1000 stems/Ha.

3

Table 3.2: Forest parameter estimation from simulation data

Parameters Real value Three-stage inversion method Proposed method

hV m

18

16.0906

17.8397

0 rad

0.0148

0.0442

0.0158

dB / m

0.2

0.3168

0.1687

RMSE

0

3.5006

2.3264

Table 3.2 shows the results of comparing the estimated forest parameters

from the proposed method with the three-stage inversion model.

Figure 3.8: Pauli image on RGB coding (a) Pauli image of the survey area,

(b) Graph comparing the forest height of two algorithms

Therefore, from Figure 3.8 (b) and Table 3.2 it could be confirmed that the

proposed algorithm has a higher accuracy than the three-stage inversion model.

The tree height in the testing forest area estimated from the proposed method

is presented in Figure 3.9. For having a deeper evaluation of the effectiveness of

the proposed method, the thesis randomly took 200 pixels in azimuth in the test

area. The main parameters of the forest are estimated by the proposed method

with 200 pixels as shown in Figure 3.10.

20

Figure 3.9: Forest height estimated by proposed approach

Figure 3.10 is a graph showing the values and standard deviations of the

forest parameters that are determined from the proposed method.

Figure 3.10: Estimate forest parameters from the proposed method

3.2.4.2. Space borne data

The used space borne data included two single-angle scans of TienShan area

by the SIR-C radar system. This is a mixed area of forests, agricultural areas and

roads. The radar system operates with with an incidence angle of 24.569 degree

and the baseline of 60m. The satellite image of the test area is shown in figure

3.11, the white rectangle was selected for the survey. The figure 3.12 (a) is an

optical image of the testing area, The figure 3.11 (b) is the Pauli decomposition

image of the testing area with 495x495 pixels.

21

Figure 3.11: The optical image of the test forest area. The white rectangle

represents the selected area for the forest evaluation

(a)

(b)

Figure 3.12: The Tien-Shan survey area a) Optical image,

b) Pauli decompositon image

Along the cutting path of the azimuth of the optical image (Figure 3.12), we

can identify three types of topography: the red areas denoting areas of bare land or

agricultural and roads, the green areas represent forest areas.

4 Figure 3.13: The compared graph of the height results

The Figure 3.13 is the histogram of the height results estimated by the

proposed algorithm compared to the three-stage inversion algorithm through 495

pixels and most of the forest height ranges from 10 m to 25 m.

The tree height in the testing forest area has an average height of 19m. The

estimated height from the proposed method shows that the forest elevation

fluctuates at an elevation of 18.18 m and is greater than the height of 16.46 m of

the three-stage inversion method shown in Figure 3.14.

22

(a)

(b)

Figure 3.14: The estimated forest height of two methods

(a) Three-stage inversion method, (b) The proposed method

5

Table 3.3: Forest parameters estimated from two approachs

Parameters

Three-stage inversion method

Proposed method

hV m

16.4633

18.1842

0 rad

-0.3110

-0.1416

dB / m

0.1043

0.1197

RMSE

4.2142

2.1814

Table 3.3 presents the results from comparing the estimated forest parameters

of the proposed method and the three-stage inversion model.

To deeper assess the effectiveness of the proposed method, the thesis

randomly took 200 pixels in azimuth in the testing area. The main parameters of

the forest are estimated by the proposed method shown in Figure 3.17. The phase

of the ground topography changed in the range of -1 to 1 rad in figure 3.17 (b), the

random anisotropic coefficient of the canopy shown in the figure 3.17 (c) and (d).

Figure 3.17: Estimate forest parameters for the survey area

23

One of the outstanding and promising applications of PolInSAR

technology has been researched and developed, that is to classify, estimate and

determine the parameters of vegetation, forest and building height.

3.4. Conclusion of Chapter 3

1. Presentation of three methods of forest height estimation with the aim of

improving the accuracy in determining forest parameters. The forest height is

estimated based on the difference of interferometric phase between the two

scattering components: it is the scattering component directly from the ground

0 and the direct scattering component from the canopy of V.

2. The proposed method of estimating surface phase 0 is based on the

coherence set, the remaining phase of the V canopy component will be

estimated from the three-component target decomposition technique for

PolInSAR images to extract the forest height.

3. Combining theoretical and empirical research. Based on the results

achieved with the simulation data, the proposed algorithm will be applied to the

experimental data received from airbonre and satellite remote sensing systems.

In addition, they could be used to directly recover other forest parameters such

as the anisotropy, the random orientation, wave attenuation in the environment.

CONCLUSION

1. The results of the thesis

The content of the thesis "Research on the accuracy improvement of the

target parameters dentification and determination using polarimetric synthetic

aperture Radar and polarimetric interferometric images" has solved the

problem of detecting and identifying targets in natural areas, urban areas and

determining forest elevation. The thesis has studied the Freeman's threecomponent scattering model and Yamaguchi's four component scattering

model, three-stage inversion algorithm and Coherence set theory. As a result,

three basic problems solved in the thesis are:

The first problem: The thesis has proposed two adaptive three-component

algorithms and four extensions with asymmetric scattering models to enhance

the capability of target identification including natural and urban areas.

The second problem: The thesis has proposed an algorithm to improve the

accuracy of forest height estimation using PolInSAR image based on the threestage inversion model, ESPRIT algorithm and coherence set.

The third problem: Three algorithms have been tested, evaluated and

compared to previous algorithms tested with the same PolSAR data set (the

observed area of Oberpfaffenhofen city of Germany) as well as simulation and

PolInSAR satellite data. The theoretical results of algorithms have been

simulated by the actual PolSAR and PolInSAR image data, the simulation

results showed the correctness of the proposed solutions and high applicability

MINISTRY OF NATIONAL DEFENCE

ACADEMY OF MILITARY SCIENCE AND TECHNOLOGY

BUI NGOC THUY

Research ON THE ACCURACY IMPROVEMENT OF THE TARGET

PARAMETERS IDENTIFICATION AND DETERMINATION USING

POLARIMETRIC SYNTHETIC APERTURE RADAR AND

POLARIMETRIC INTERFEROMETRIC IMAGES

Specialization: Electronic engineering

Code:

9 52 02 03

SUMMARY OF DOCTORAL THESIS IN ENGINEERING

Hanoi, 2019

This thesis has been completed at:

ACADEMY OF MILITARY SCIENCE AND TECHNOLOGY

Scientific Supervisors:

1. Assoc. Prof, Ph.D Le Vinh Ha

2. Ph.D Pham Minh Nghia

Reviewer 1: Prof, Ph.D Bach Gia Duong

University of Engineering and Technology, Vietnam

National University, Hanoi

Reviewer 2: Assoc. Prof, Ph.D Hoang Van Phuc

Military Technical Academy

Reviewer 3: Ph.D Le Thanh Hai

Academy of Military Science and Technology

This thesis was defended at the Doctoral Evaluating Council at Academy

level held at Academy of Military Science and Technology

in 8:30, date ….month ….. year 2019.

The thesis can be found at:

- Library of Academy of Military Science and Technology

- Vietnam National Library

1

INTRODUCTION

1. The urgency of thesis

Over the last two decades, with the development of science and technology,

remote sensing data has been used more widely in many fields. Remote sensing

applications have been supporting effectively in tactics and modern warfare by

creating the unexpected elements, surveillance and locate targets exactly, to name

but a few. By processing remote sensing data that are obtained from satellites and

aeronautics with very high resolution, the identification and determined of objects

such as military targets and civilian targets could provide the relative accuracy

without making physical contact with the objects.

Particularly, in order to assess the impact of changing forest ecosystems, ice,

as well as the phenomenon of global climate change, remote sensing technology

has been researched and developed and applied in varied fields of socio-economic

life. Technique of Polarimetric Synthetic Aperture Radar and Polarimetric

Interferometric (PolSAR and PolInSAR) is a topic that is drawing the attention of

scientists all over the world. This technology contributes to deal with the problems

of the object‘s identification and measuring target parameters. Therefore, the

thesis has its scientific and practical significance.

Therefore, the PhD student has chosen the topic: Research on the accuracy

improvement of the target parameters identification and determination using

polarimetric synthetic aperture Radar and polarimetric interferometric images.

2. The objective

Research scientific basis and solutions to improve the accuracy of

identification and identify parameters of natural targets, targets and forest height

estimates based on PolSAR and PolInSAR images.

3. The subject and scopes

PolSAR's target decomposition technique is based on Freeman's threecomponent scattering model and Yamaguchi's four-component scattering model.

The PolInSAR decomposition technique is based on the coherence set and the

three-component decompositon technique for PolInSAR images. From above

researchs, proposing and developing solutions and simulations for the accuracy

improvement of interpretation, identification and determination of target

parameters are studied by the thesis.

4. Research methodology

- Methods of collecting information, documents, general analysis of scientific

works and articles published in the world and in the country. Collect PolSAR and

PolInSAR image data sources related to the test area.

- Researching target decomposition techniques and building algorithm models

to improve the accuracy of identification and determination of target parameters

based on PolSAR and PolInSAR images.

- Programming technology and application of informatics technology in

building a program to perform calculations and simulations using MATLAB tool

2

combined with specialized software ENVI 5.0 and PolSARproSim, performing

verification experiments.

5. Scientific and practical significance of the thesis

- The research results of the thesis have contributed to the decomposition

technical theory based on the three-component scattering model and four

components for PolSAR image with asymmetric scattering model to improve the

ability to identify natural objects, artificial targets and to improve the accuracy of

estimating forest height in PolInSAR images.

- The research results of the thesis provide a full assessment of the scientific

basis as well as a test result of solutions to improve the identification accuracy and

identify targets that can be applied for military and civil purposes. Serving

teaching research, specialized research, realizing applied software and developing

remote sensing technology in Vietnam.

6. Structure of the thesis

The thesis is composed of the beginning, 3 chapters, the conclusion is as

follows:

Chapter 1: Overview of identification and determination of target’s

parameters based on polarized radar and polarization-interference images; Chapter

2: Propose target identification algorithm based on PolSAR image; Chapter 3:

Estimating target parameters based on PolInSAR images; Finally, the conclusions,

assessments and the issues that need further research.

Chapter 1. OVERVIEW OF IDENTIFICATION, DETERMINATION

OF TARGET PARAMETER BASED ON POLARIZED RADAR

AND POLARIZATION INTERFERENCE IMAGES

1.1. Radar equation

The electromagnetic waves of radar system transmitted in the medium can be

obtained at a target and the energy carried by the incident wave is absorbed by the

target itself, whereas the rest is reradiated as a new electromagnetic waves. Due to

the interaction with the target, the properties of the reradiated waves can be

different from those of the incident ones. From this change, they can help us

describe or identify targets. In reality, we are interested in the changes concerning

the polarization of the wave.

Figure 1.1: The interaction of electromagnetic waves with a target

3

The radar equation represents as the following:

P G ( , ) A ( , )

(1.1)

PR T T 2 ER 2

4 rT

4 rR

1.2. The formation of Synthetic Aperture Radar image

SAR image systems allow for monitoring the earth on a global scale at all

times and in all weather conditions. The basic geometry of a SAR system is shown

in Figure 1.3.

Figure 1.3. Basic geometric

Figure 1.4. Ground range to slant

structure of a SAR space

range projection

One of the most important criteria for assessing the quality of SAR image

systems is its spatial resolution. The spatial resolution describes the visibility of

the image as much as possible so that two scattered objects can be separated in

term of spatial meaning.

The imaging SAR system is a side-looking radar sensor with an illumination

perpendicular to the flight line direction as shown in Figure 1.4.

1.3. Characteristics of target polarization.

1.3.1. Polarized information

Polarized information in backscattered waves from a given environment can

be related to: reflecting geometric structures such as shape and orientation or

geophysical structure such as moisture, surface roughness face…

(a) Polarimetric SAR (PolSAR)

(b) Single- polarization SAR

Figure 1.8: Polarizing types in remote sensing radar

4

1.3.2. Target scattering vector k

The relationship between incident and scattered waves is as follows [42]:

ES

e jkr

e jkr S11

S EI

r

r S21

S12

EI

S22

(1.12)

To be able to extract physical information from the 2×2 coherent

backscattering matrix or Sinclair matrix which achieved through the construction

of system vectors, we represent the Sinclair matrix with the vector V(.) as follows:

S

S HH

SVH

S HV

SVV

1

k V S Trace S

2

(1.13)

Where ψ is a set of 2×2 complex basic matrices which are constructed as an

orthogonal set under the Hermitian inner product.

1.3.3. Coherency matrix [T] and covariance matrix [C] polarimetric

The polarimetric Pauli coherency matrix [T] and the Lexicographic

covariance matrix [C] are generated from outer product of the associated target

vector with its conjugate transpose as.

T

k3P .k 3*TP

and C k3L .k 3*TL

(1.34)

Where superscript T and * denote transpose and complex conjugation and

denotes the ensemble average in the data processing, respectively.

1.4. Scattering mechanisms

- Surface scattering.

- Double-bounce scattering.

- Volume scattering.

- Helix scattering.

- Wire scattering.

1.5. Target polarimetric characterization

Symmetric scattering hypotheses about the distribution of scattering objects

will make scattering problems simple and allow for quantitative conclusions about

their scattering properties.

1.6. Polarimetric target decomposition PolSAR

Model-based decompositions for the target based on physical scattering

models to interpret the scattering process.

(1.67)

C CS CD CV

In addition, the total power P of all components can be retrieved by

summation of the power contributions PS, PD, PV, which are calculated from the

trace of the single scattering component matrices.

(1.68)

P Tr CS Tr CD Tr CV PS PD PV

Finally, the normalized power of each scattering component can be obtained

by division with P. This provides the opportunity to compare the strength of the

different scattering contributions with respect to each other.

5

1.7. Interferometry Synthetic Aperture Radar

Interferential Synthetic Aperture Radar (InSAR) takes advantage of the phase

difference between two SAR complex images described from differences at

locations or at different times.

Figure 1.21: Geometric structure of interference radar

When building the complex product s1s2* for interferometry, we could cancel the

scattering phase terms and keep the geometrical phase. In effect, we now obtain a

signal phase which depends only on the difference in range between the two positions

R R1 R 2 as follows:

j 2 2 R1s1

4

j

R

s1 a1e

*

(1.71)

s

s

Ae

1

2

2

j 2 R2 s2

s2 a2 e

The observed component in these interferometric radar systems is the

interference phase and is determined as follows:

4

(1.72)

arg s1s2*

R 2 N ; N 0, 1, 2,...

1.8. Polarimetric Interferometric Synthetic Aperture Radar

PolInSAR differs from conventional SAR interferometry in that it allows

generation of arbitrary interferogram and receive polarization pairs as Figure 1.22.

Using the outer product formed from the scattering vectors k1 and k2 for

images S1 and S2, we can define a 6×6 Hermitian positive semidefinite matrix [T6]

and [C6] as follow:

k

3 P1 k *T

T6

3P1

k

3P2

k3*T

P

k

3L1 k *T

C6

3L1

k

3L2

k3*T

L

T1

2

*T

2

C1

C *T

int

T2

Cint

C2

(1.76)

(1.77)

6

The interference equation is as follow:

arg 1k3P k 3*TP *2T

1

2

arg

*T

1

2

(1.80)

By using Eq. (1.85), the complex interferometric coherence as a function of

the polarization of the two images may be written as:

12*

1*2

1, 2

(1.81)

11* 2 2*

1*T T1 1 *2T T2 2

Figure 1.22: PolInSAR acquisition geometry

1.9. Conclusion of Chapter 1

Through an overview of the target decomposition technique based on PolSAR

and PolInSAR images, the works have been published in foreign and domestic

scientific journals as well as factors affecting the problem of identification and

correctly determination the target parameters shows:

- Research on resolving disadvantages in pixels still has many negative power

components, this is the main cause leading to inaccurate identification of targets.

- Volume scattering components are often assumed to be symmetric reflective

scattering, the number of observations is limited because some assumptions must be

accepted to eliminate ambiguity, assumptions often cause the negative power in the

scattering mechanisms leads to incorrect assessment of forest height.

Therefore, the thesis is to set out 2 main problems to solve:

Building PolSAR image processing algorithms based on scattering models in

response to the requirements of interpretation capacity enhancement, target

parameters determination such as the accuracy of natural and artificial targets

identification. This problem is solved in chapter 2.

A proposed algorithm to improve the accuracy of forest height estimation

based on simulation and satellite data sources. This problem is solved in chapter 3.

7

1

Chapter 2. A PROPOSED ALGORITHM FOR IDENTIFYING

TARGET BASED ON POLSAR IMAGE

2.1 Targeted decomposition techniques based on polarized radar images

2.1.1. Technical analysis of coherence target

The scattering model after rotating an angle can be expressed as simple

conversion as in the following expression (2.1):

sin

cos sin SHH SHV cos

S

(2.1)

cos SVH SVV sin cos

sin

Using a unitary transformation matrix according to the parameters of the Pauli

rotation matrix, we can express them in a form of unitary vector as follows:

1

1 0

k

2 0

0

0

cos 2

sin2

0

0

sin2

cos 2

0

0 S HH SVV

0 S HH SVV

0 S HV SVH

1 S HV SVH

(2.2)

We find that the complex totals SHH SVV and SHV SVH are invariable for the

axis of rotation, which gives them a special physical meaning.

2.1.2. Technical analysis of targets according to the scattering model

The decomposition technique following the basic scattering model as shown

in Figure 2.2 including scattering of the surface, double bounce and volume

scattering:

Figure 2.2: Basic scattering models

2.1.2.1. Freeman-Durden three component decomposition

The Freeman-Durden (FDD) decomposition is a technique for fitting a

physically based, three component scattering mechanism model to the PolSAR

observations, without utilizing any ground truth measurement.

The Freeman-Durden decomposition model:

C3V C3S C3 D C3V

f 2 f 2 3 fV

D

S

8

0

f * f * fV

D

S

8

0

2 fV 8

0

8

0

3 fV

fS fD

8

f S f D

fV

(2.17)

8

Equation (2.17) gives us four equations with five unknowns. However, the

contribution of a number

fV 2 fV

3f

,

or V of volume scattering can be eliminated

8 8

8

2

2

*

, then we get 3 equations in 4 unknown equations:

SHH , SVV and S HH SVV

*

S HH S HH

fS fD

2

2

*

S HH SVV

f S f D

(2.18)

*

SVV SVV

fS fD

The fS, fD coefficients and or parameters can be used to determine the

properties of the target. Finally, the contribution of each scattering mechanism can

be determined for the following spans:

Span SHH 2 SHV SVV PS PD PV

2

2

PS f S 1

With:

2

;

2

PD f D 1

2

;

(2.19)

PV fV

(2.20)

2.1.2.2. Yamaguchi decomposition

The three-component scattering model using a covariance matrix is very

effective for scattering mechanisms in PolSAR and the algorithm is only based on

*

*

symmetric reflections SHH SHV SHV SVV 0 , which is the limitation of the FDD

*

*

technique [12], due to SHH SHV 0, SHV SVV 0 in reality.

The Yamaguchi decomposition model:

C f S CS f DCD fV CV f H CH

2

fS 0

*

0

0

0 d

0

0

b

0

2

0 fD 0

*

1

0

a

0 fV 0

1

d

0

1

0 fH j 2

c

1

j 2

2

2*

1

j 2

(2.28)

1

In which, fS, fD, fV, fH indicate the determined scattering coefficients,

corresponding to the surface scattering, double bounce, volume and helix

scattering mechanisms.

The contribution of each scattering mechanism can be estimated as:

2

2

PS f S 1 ; PD f D 1 ; PV fV a b c ; PH f H

(2.29)

2

2

2

Pt PS PD PV PH SHH 2 SHV SVV

2.2. Target identification based on the three-component scattering

decomposition with an adaptive volume modelling

Adaptive algorithm solves the problem of general eigenvalues analysis and

non-negative power constraints, thereby determining the unique minimum value

for volume scattering. Finally, we determine the power of two remaining

scattering components. The generalized double-bounce scattering component

reflects the interaction of electromagnetic waves in both natural and urban areas.

9

* Proposed decomposition method:

The coherence matrix of target is analyzed into a combination corresponding

to a practical scattering mechanism.

Then [T] by given as

1

PS

T f S TS f DTD fV TV

2

1

0

*

0

2

0

P

0 D

0

1

*

2

*

*

a d 0

P

* V d b 0 (2.40)

2

0 0 c

2

=1+ ; a b c

(2.41)

With , , are parameters of the surface scattering and double bounce

scattering models

When is known, we can calculate the remaining power Pv, and PS.

.P 2

P P

2

PD Pt ; PV (T33 t . ); PS 1 T11 D V a

(2.49)

c

Algorithm flowchart:

2

2

The coherence matrix [T]

Determined TS, TD, TV

0.3

PV 0; PD .Pt

PS 1

2

T

11

*

T23

T33

;

PD

T33

*

;

PD 0; PV c

PS 1

2

T

11

PV

0; 0;

T13

PD 2

P

T11 D

T22

PV

b

P

T11 V a

T22

Calculate PS., PD and PV using parameters , and

Figure 2.5: The flowchart of proposed algorithm

To evaluate the proposed decomposition method, the experiments were

performed using the full PolSAR data of ESAR (Experimental Synthetic Aperture

Radar) with 3x3m resolution. By testing the area near Oberpfaffenhofen,

10

Germany, Figure 2.6 (a) shows the optical image of this area. The is the ratio of

asymmetric scattering power to total power. Figure 2.6 (c) shows the color image

of the coefficient.

Figure 2.6: Survey area, (a) Optical image of Oberpfaffenhofen area,

(b) Pauli image, (c) Image color of correlation coefficient

A

B

The proposed

decomposition

(a)

(b)

Freeman decomposition

Figure 2.7: Decomposition image of the test area, (a) Color image of the three

components of the proposed decomposition method, (b) Color image of the three

components of the Freeman decomposition method

* Experimental results:

The effectiveness of the proposed method is evaluated in comparison with

Freeman's three-component decomposition. From the compared results, we can

evaluate accurately with dominant double bounce scattering mechanisms in the

urban area and dominant volume scattering in the forest area. In the figure 2.7 (a),

the volume scattering component determined from the proposed decomposition

shows that the observation is clearer than that in the Freeman decomposition

technique as shown in Figure 2.7 (b). In Figure 2.7 (b) many pixels in the image of

the urban area are still green, thus lead to misinterpret and misidentify the target.

The main reason is the existence of many pixels with negative power components

in the Freeman decomposition technique.

11

In Figure 2.8, we could find that the value is quite low corresponding to

forest area and agricultural land. However, in areas with artificial structures or

urban areas, the value is quite large.

Figure 2.8: Graph at the test areas (a) forest area,

(b) farmland area, (c) urban area.

Thus, the proposed algorithm has added the asymmetric double bounce

scattering component, consequently, the image results have significantly

improved, and the target identification and determination are more precicely than

that in the Freeman decomposition.

Forest

Agricultural land

Urban

The proposed method

(a)

(d)

Surface scattering

(b)

Freeman decomposition

(e)

Double bounce scattering

(c)

(f)

Volume scattering

Figure 2.9: Pie chart of three components of scattering in surveyed areas.

(a,b,c) the proposed decomposition method, (d,e,f) Freeman decomposition

12

It could be seen that the stability of the proposed decomposition method

compared to the Freeman dicomposition method is more clearly in comparing the

ratio of pixels with non-negative power components to filtering windows of

different sizes, shown in the Table 2.1.

Table 2.1: The table of the ratio of pixels with non-negative power components

Filter window size

1×1

3×3

7×7

9×9

Proposed method

99.09%

99.14%

99.20%

99.32%

Freeman method

97.05%

98.21%

98.64%

98.83%

From the above results, comparisons and analysis, the proposed method is

better and more stable than Freeman's three-component decomposition method.

Although asymmetric double bounce scattering could not completely solve all the

scattering problems in urban areas. Some pixels of buildings and the transport

system are still misinterpreted and misidentification.

2.3. Urban target identification based on four-component decomposition

technique with extended volume scattering model

In the proposed method, the thesis uses adaptive algorithms to determine the

covariance matrix parameters of volume scattering and asymmetric scattering. The

effectiveness of the proposed method is proved with PolSAR data obtained from

ESAR airborne remote sensing system.

* Proposed decomposition method:

To better describe asymmetric scattering, the method will take into account

all parameters related to asymmetric scattering information, the proposed

decomposition model is presented as follows:

C f S CS f DCD fV CV f asymCasym

2

f S 0

*

0

0

0

2

0 f D 0

*

1

0

0

0

a

0 fV e*

*

1

d

e

b

f*

d

f

c

f asym

2

2

*

2 *

*

2

2

2 *

2

1

(2.54)

2

2

where 2 1; a b c; Pasym is the asymmetric scattering

power component; 2 , 2 are both complex numbers corresponding to C12

and C23, The scattering terms can be utilized to describe the general case of

nonreflection symmetric scattering, generally, C12 0 and C23 0 and a f are

the elements in the volume scattering model.

Compared to the Yamaguchi four component decomposition method, the

proposed decomposition method uses the asymmetric scattering mechannism.

Therefore, it can be seen that the asymmetric scattering often appears in complex

urban areas, and disappears in natural distributed areas.

* Experimental results:

The effect of the proposed algorithm is evaluated based on dataset received

from the E-SAR airbonre system and the test area which is close to

Oberpfaffenhofen, Germany.

13

Figure 2.11: Optical image of the survey area (a) Optical image GoogleEarth,

(b) Image color of correlation coefficient

A

B

C

The proposed decomposition

(a)

(b)

Yamaguchi decomposition

Figure 2.12: Image decomposition of the test area (a) The proposed

four-component method, (b) Yamaguchi decomposition

From the comparison of image results shown in Figure 2.12 (a) we could see

that urban areas are purplish red, better than those in Yamaguchi's decomposition

in Figure 2.12 (b). From the results of the proposed decompositon method, the

target identification and determination of the proposed method are more precise

than that in the Yamaguchi decomposition technique.

The proposed method utilise an asymmetric scattering component in order to

improve the limitation of Yamaguchi decomposition in terms of the interpretation,

estimation, and classification of the terrain of the targets. From the above results,

comparisons and analysis, the proposed method shows better improvement of

Yamaguchi's four-component decomposition method. Although the proposed

method still has the disadvantage is the forest areas located bottom right image of

Figure 2.12 (a) misidentification due to asymmetric scattering component exceeds

the power level. In the future, the thesis will continue to research further

improvement and hope for better results aimed at improving the effectiveness of

the proposed method.

14

Forest

Agricultural land

The proposed method

Urban

(a)

(b)

Yamaguchi decomposition

(c)

(d)

(e)

(f)

Surface scattering

Double bounce scattering

Volume scattering

Asymmetric scattering

Figure 2.13: Pie chart of four-components of scattering in surveyed areas.

(a,b,c) the proposed decomposition method, (d,e,f) Yamaguchi decomposition

2.4. Conclusion of Chapter 2

1. The overview of the Freeman and Yamaguchi target decomposition

techniques.

2. Build two target identification models based on three-component and fourcomponent decomposition techniques with asymmetric scattering models.

3. The program is set up in Matlab environment. Then, based on the results

extracted from Matlab software, the thesis continues to use specialized software

ENVI 5.0 to perform interpretation and analysis of target.

4. Experimental results of the two proposed algorithms at the points with a

negative power component significantly reduced. Therefore, there has been a

significant improvement in the identification and interpretation of targets more

accurately than the Freeman decomposition technique and the Yamaguchi

decomposition technique.

From the results of the above study, the proposed methods show improved

accuracy in interpretation, identification and determination of target.

Chapter 3. ESTIMATING PARAMETERS OF TARGET

BASED ON POLINSAR IMAGE

3.1. Methods of height estimating of target

3.1.1. Three-stage inversion technique

Forest height estimation in three-stage inversion technique can be divided

into three separate stages.

15

The implementation of the three-stage inversion method is summarized in

Figure 3.1.

Figure 3.1: Three-stage inversion method

The forest height estimation is usually determined after eliminating the

phase component of the surface in complex interferometric coherence coefficients

of the volume scattering component in terms of an inversion space in the unit

circle. A vol value in the inversion space can be used to recover the pair of value

of height and extintion coefficient.

The definition of average height of tree is:

1 N

hV hV i

(3.2)

N i1

The root mean square error is defined as:

N

RMSE

hV i hV

i 1

N

2

(3.4)

Where hV-i is the tree height estimation at i-th pixel, and N is the number of

samples used for the calculation.

3.1.2. Estimation of Signal Parameters via Rotational Invariance Techniques

The main idea of this method is to transform the observation space into two

interpolated subspace: signal and noise. ESPRIT algorithms can be extended to

analyze SAR data. For SAR data, the backscattering signal received from the antenna

is analyzed into a total of different scattering mechanisms. The interferometry phases

depend strongly on the observed medium.

16

2 Figure 3.2: ESPRIT algorithm

3.1.3. Coherence set theory

The continuous polarization space enables us to use the polarization diversity

and to consider the set of all coherences as a continuous entity. We will refer to

the set of all coherences in the following as the coherence set. It has been

identified that the coherence set is directly connected with the mathematical

concept of the numerical range (also known as field of values) of the contraction

matrix. This section is aim to present some theoretical aspects of the PolInSAR

coherence set. The first part of this section deals with the different definitions of

the numerical range. The second and third parts present principles for information

extraction of the coherence set geometry and the coherent properties in the

complex coherence plane.

3.2. The forest parameter estimation based on the target decomposition

technique for PolInSAR images

In the proposed method, the accuracy of forest parameter estimation can be

improved by a coherence method. First, the phase and the parameters of the

scattering object in the tree canopy are determined by the target analysis algorithm

by the adaptive scattering model, then removing the scattering component from

the covariance matrix and the phase of the terrain is determined by the ESPRIT

algorithm. Finally, the forest height is estimated and compensated by the

coherence compensation method. Experimental results show that the accuracy of

forest parameter estimation is significantly improved.

3.2.1. Polarization interference coherence

The data received from the PolInSAR system are usually represented by a

complex coherence matrix 6×6, and are represented in (3.28).

17

T kk

T1

*T

*T

T2

k

1

k 2

với k

(3.28)

The polarization interference coherence of PolInSAR system could be

described by a polarization function of two images and represented as follows:

1*T 2

*T

(3.29)

1, 2

*T

1*T T1 1 *2T T2 2 T

where 1 2 is a unit complex vector of each polarization channel,

T T1 T2 2 . The contracted form of coherence matrix for PolInSAR data is given by:

T

1 2

T

1 2

11 12

21 22

0

0

0

0

33

(3.30)

3.2.2. Scattering mechanisms for PolInSAR data

- Surface scattering

- Double-bounce scattering

- Volume scattering

3.2.3. Estimating scattering parameters of trees

The contracted coherence matrix PolInSAR's is analyzed into the sum of three

sub-matrices corresponding to the three scattering components: volume scattering,

double bounce scattering and surface scattering:

j

f S e S TS f D e jD TD fV e jV TV

(3.49)

where f S , f D .and fV represent the scattering power coefficient of single bounce,

double bounce and volume scattering respectively.

If 11 < 33 then the parameters of scattering components are determined

as follows:

fV 1

2

11

; V arg 1

2

11

2

2

0 ; f S 0 ; f D 22 33 11 ; D arg 22 33 11

(3.50)

Conversely, if 11 33 , then parameters of random scattering

components from tree canopy are directly determined from (3.49)

fV 33 ; V arg 33

T

TV 33

v33

FS f S e jS

FD f D e

jD

jV

; FV fV e ; arcos

11 22 TV 11 TV 22 FV

11

12

;

1

22 33 FD

22 TV 11 TV 22 FV

2

4 12 TV 12 FV

2

2

11 22 TV 11 TV 22 FV

11 22

2

TV 11 TV 22 FV

2

4 12 TV 12 FV

2

(3.51)

18

3.2.3. Estimate the parameters of terrain using the coherence set

In order to reduce the computational complexity and to improve the

accuracy of the terrain parameter estimation, the thesis proposes the using of the

coherence set to directly estimate the parameters of the terrain.

Algorithm flowchart:

PolSAR

Dataset 1

PolSAR

Dataset 2

Co-register photos

Contraction matrix

Estimating scattering parameters of

trees by 3-component technique

11 33

V arg 1

Estimating surface scattering

parameters using the coherence set

11 33

33

T

V 33

2

11

hV

V arg

V 0

kZ

V 0

0 arg 2 31 L

R sin

4 Bcos

Figure 3.7: The flowchart of proposed algorithm

Based on the characteristics of contracted coherence matrix in (3.31), we have

a coherence set for PolInSAR data as follows:

(3.52)

app w*T w : w*T w 1, w 3

Equation (3.52) has the same form as the numerical range of the square

3

matrix A . Therefore, the numerical range of the matrix can also be

considered as the region of the coherence set.

The interferometric phase of the ground topography is determined as:

0 arg 2 3 1 L

(3.53)

Finally, the forest height is determined by the differency between the

scattering phase of the canopy and the scattering phase of the ground topography,

as shown in (3.56).

0

R sin

(3.56)

hV V

V 0

kZ

4 B cos

19

Where is the angle between the radiation wave and the vertical axis, R

is the distance between radar and target, is the angle of deviation between the

baseline and the horizontal axis, is the electromagnetic wave.

3.2.4. Experimental results

3.2.4.1. Simulation data

The simulated data obtained from PolInSAR system L-band (1.3 GHz)

with 100m horizontal and 10m vertical baseline is shown in Figure 3.8. The test

forest area has an average height of 18m on a relatively flat topography with the

forest stands occupies a 2.6 Ha area and stand density is 1000 stems/Ha.

3

Table 3.2: Forest parameter estimation from simulation data

Parameters Real value Three-stage inversion method Proposed method

hV m

18

16.0906

17.8397

0 rad

0.0148

0.0442

0.0158

dB / m

0.2

0.3168

0.1687

RMSE

0

3.5006

2.3264

Table 3.2 shows the results of comparing the estimated forest parameters

from the proposed method with the three-stage inversion model.

Figure 3.8: Pauli image on RGB coding (a) Pauli image of the survey area,

(b) Graph comparing the forest height of two algorithms

Therefore, from Figure 3.8 (b) and Table 3.2 it could be confirmed that the

proposed algorithm has a higher accuracy than the three-stage inversion model.

The tree height in the testing forest area estimated from the proposed method

is presented in Figure 3.9. For having a deeper evaluation of the effectiveness of

the proposed method, the thesis randomly took 200 pixels in azimuth in the test

area. The main parameters of the forest are estimated by the proposed method

with 200 pixels as shown in Figure 3.10.

20

Figure 3.9: Forest height estimated by proposed approach

Figure 3.10 is a graph showing the values and standard deviations of the

forest parameters that are determined from the proposed method.

Figure 3.10: Estimate forest parameters from the proposed method

3.2.4.2. Space borne data

The used space borne data included two single-angle scans of TienShan area

by the SIR-C radar system. This is a mixed area of forests, agricultural areas and

roads. The radar system operates with with an incidence angle of 24.569 degree

and the baseline of 60m. The satellite image of the test area is shown in figure

3.11, the white rectangle was selected for the survey. The figure 3.12 (a) is an

optical image of the testing area, The figure 3.11 (b) is the Pauli decomposition

image of the testing area with 495x495 pixels.

21

Figure 3.11: The optical image of the test forest area. The white rectangle

represents the selected area for the forest evaluation

(a)

(b)

Figure 3.12: The Tien-Shan survey area a) Optical image,

b) Pauli decompositon image

Along the cutting path of the azimuth of the optical image (Figure 3.12), we

can identify three types of topography: the red areas denoting areas of bare land or

agricultural and roads, the green areas represent forest areas.

4 Figure 3.13: The compared graph of the height results

The Figure 3.13 is the histogram of the height results estimated by the

proposed algorithm compared to the three-stage inversion algorithm through 495

pixels and most of the forest height ranges from 10 m to 25 m.

The tree height in the testing forest area has an average height of 19m. The

estimated height from the proposed method shows that the forest elevation

fluctuates at an elevation of 18.18 m and is greater than the height of 16.46 m of

the three-stage inversion method shown in Figure 3.14.

22

(a)

(b)

Figure 3.14: The estimated forest height of two methods

(a) Three-stage inversion method, (b) The proposed method

5

Table 3.3: Forest parameters estimated from two approachs

Parameters

Three-stage inversion method

Proposed method

hV m

16.4633

18.1842

0 rad

-0.3110

-0.1416

dB / m

0.1043

0.1197

RMSE

4.2142

2.1814

Table 3.3 presents the results from comparing the estimated forest parameters

of the proposed method and the three-stage inversion model.

To deeper assess the effectiveness of the proposed method, the thesis

randomly took 200 pixels in azimuth in the testing area. The main parameters of

the forest are estimated by the proposed method shown in Figure 3.17. The phase

of the ground topography changed in the range of -1 to 1 rad in figure 3.17 (b), the

random anisotropic coefficient of the canopy shown in the figure 3.17 (c) and (d).

Figure 3.17: Estimate forest parameters for the survey area

23

One of the outstanding and promising applications of PolInSAR

technology has been researched and developed, that is to classify, estimate and

determine the parameters of vegetation, forest and building height.

3.4. Conclusion of Chapter 3

1. Presentation of three methods of forest height estimation with the aim of

improving the accuracy in determining forest parameters. The forest height is

estimated based on the difference of interferometric phase between the two

scattering components: it is the scattering component directly from the ground

0 and the direct scattering component from the canopy of V.

2. The proposed method of estimating surface phase 0 is based on the

coherence set, the remaining phase of the V canopy component will be

estimated from the three-component target decomposition technique for

PolInSAR images to extract the forest height.

3. Combining theoretical and empirical research. Based on the results

achieved with the simulation data, the proposed algorithm will be applied to the

experimental data received from airbonre and satellite remote sensing systems.

In addition, they could be used to directly recover other forest parameters such

as the anisotropy, the random orientation, wave attenuation in the environment.

CONCLUSION

1. The results of the thesis

The content of the thesis "Research on the accuracy improvement of the

target parameters dentification and determination using polarimetric synthetic

aperture Radar and polarimetric interferometric images" has solved the

problem of detecting and identifying targets in natural areas, urban areas and

determining forest elevation. The thesis has studied the Freeman's threecomponent scattering model and Yamaguchi's four component scattering

model, three-stage inversion algorithm and Coherence set theory. As a result,

three basic problems solved in the thesis are:

The first problem: The thesis has proposed two adaptive three-component

algorithms and four extensions with asymmetric scattering models to enhance

the capability of target identification including natural and urban areas.

The second problem: The thesis has proposed an algorithm to improve the

accuracy of forest height estimation using PolInSAR image based on the threestage inversion model, ESPRIT algorithm and coherence set.

The third problem: Three algorithms have been tested, evaluated and

compared to previous algorithms tested with the same PolSAR data set (the

observed area of Oberpfaffenhofen city of Germany) as well as simulation and

PolInSAR satellite data. The theoretical results of algorithms have been

simulated by the actual PolSAR and PolInSAR image data, the simulation

results showed the correctness of the proposed solutions and high applicability

## Nghiên cứu các giải pháp nâng cao độ chính xác đo cao GPS trong điều kiện Việt Nam

## các giải pháp nâng cao độ chính xác đo cao GPS trong điều kiện việt nam

## Đề tài: "Nghiên cứu các giải pháp nâng cao độ chính xác đo cao GPS" pptx

## Nghiên cứu giải pháp nâng cao hiệu quả công trình thủy lời các tỉnh trung du miền núi phía bắc phục vụ đa mục tiêu

## NGHIÊN CỨU GIẢI PHÁP NÂNG CAO ĐỘ AN TOÀN BẢO MẬT CHO HỆ THỐNG THÔNG TIN ỨNG DỤNG TẠI TRƯỜNG CAO ĐẲNG CỘNG ĐỒNG HÀ TÂY

## Nghiên cứu các giải pháp nâng cao độ chính xác đo cao GPS trong điều kiện Việt Nam

## Nghiên cứu phương pháp nâng cao độ chính xác chuyển trục chính công trình lên sàn tầng xây dựng

## Nghiên cứu giải pháp nâng cao độ bền cho bê tông – bê tông cốt thép của kết cấu bảo vệ mái đê và bờ biển Việt Nam

## NGHIÊN CỨU GIẢI PHÁP CẢI THIỆN ĐỘ CHÍNH XÁC CHO CÁC BỘ THU HỆ THỐNG ĐỊNH VỊ SỬ DỤNG VỆ TINH KHI HOẠT ĐỘNG TRONG ĐIỀU KIỆN MÔI TRƯỜNG PHỨC TẠP

## Nghiên cứu giải pháp nâng cao độ chính xác vị trí đổ sợi của hệ thống đổ sợi tự động máy sợi con (tt)

Tài liệu liên quan