ASAR APP & APM Image Quality Peter Meadows & Trish Wright

Page 1ASAR Validation Review - ESRIN – 11-12 December 2002

Advanced Technology Centre

ASAR APP & APM Image Quality

Peter Meadows & Trish Wright Properties of APP & APM Products

Example APP & APM Products

Analysis Approach

Format Verification

Visual Inspection

Impulse Response Function Measurements

AP Cross-Polarisation Ratio

AP Channel Co-Registration

Equivalent Number of Looks and Radiometric Resolution

Azimuth Ambiguities

Localisation Accuracy

Preliminary Radiometric Calibration

Noise Equivalent Radar Cross-Section

Summary



Properties of APP Imagery

• Ground range detected alternating polarisation imagery• Dual polarisation (HH & VV, HH and HV or VV & VH)• Elevation antenna pattern and range spreading loss corrections

applied• Size up to 300Mbytes with 2 byte (16bit) amplitude pixel values• Swath widths of 100 km (IS1) to 56 km (IS7) with azimuth

extents of ~100 km• Azimuth resolution of 27.6m (2 looks of ~250 Hz each)• Range resolution from 21m (IS2 far range) to 37m (IS1 near

range) and ~26m for IS3 to IS7 (1 look)• 12.5m by 12.5m pixels (hence under-sampling for spatial

resolutions less than 25m)



Properties of APM Imagery

• Ground range detected alternating polarisation medium resolution imagery

• Dual polarisation (HH & VV, HH and HV or VV & VH)• Elevation antenna pattern and range spreading loss corrections

applied• Size up to few Mbytes with 2 byte (16bit) amplitude pixel values• Swath widths of 100 km (IS1) to 56 km (IS7) with azimuth

extents of ~100 km up to 4000 km• Azimuth resolution of 135m (10 looks of ~50 Hz each)• Range resolution from 109m (IS1 far range) to 163m (IS1 near

range) and ~130m for IS3 to IS7 (1 look)• 75m by 75m pixels



APP & APM Products

IS1 IS2 IS3 IS4 IS5 IS6 IS73590 3633 3898 3855 3583 35733579

IS1 IS2 IS3 IS4 IS5 IS6 IS73712 3669 3812

AP HH/HV

AP VV/HV

• The blue orbit numbers are of The Netherlands and the purple are of Resolute, Canada. All products are processed with v3.03

• No IS1 data• Two APP products had saturated ASAR transponders (orbits 3812 & 3855)



Example APP Products

14 July 2002

IS2 HV & HH

Ottawa, Canada



6 November 2002

IS2 HV & HH

Resolute, Canada



6 November 2002

IS2 HV & HH

The Netherlands



9 November

2002

IS3 HV & HH

Example APM Products

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

0 20 40 60 80

Ground Range (km)

Rela

tive S

igm

a0 (

dB

)



12 November 2002 IS3 VV & VH



Analysis Approach• Format verification using EnviView & the ESA SAR Product

Control Software (developed by DLR & BAE Systems)• Image analysis performed using SAR Control Software (point

target & calibration modules) and following the ESA document on Quality Measurement Definitions for ASAR products

Format Verification• No problems identified with APP & APM format or header parameters,

however many of the APP products had no Chirp Parameter ADSR (hence no annotated Chirp Powers)

Visual Inspection• No problems found with any post Orbit 3661 v3.03 APP or APM products



Impulse Response Function Measurements (APP)• IRF parameters have been derived using the ESA transponders in The

Netherlands and the Radarsat transponders in Canada (images 1.6 by 1.6 km).

Fredericton (HH)

Resolute (HH)

Aalsmeer (HV)

Edam (HH)

Swifterbant (HV) Resolute (HV)

Fredericton (HV)Edam (HV)

Zwolle (VH)

Edam (VV)



Example APP Impulse Response Functions

Radarsat transponder, i = 15.81°, azimuth resolution = 27.49m, range resolution = 34.70m

ASAR transponder, i = 23.85°, azimuth resolution = 27.83m, range resolution = 24.14m



Spatial Resolution (3dB width of IRF):

Note that the IS1 swath extends to 22.2° and that the IS2 swath starts at 19.2°.

• Azimuth resolution (y): 28.09±1.97m (c.f. ~27.6m theoretical value, +10% limit & 30m requirement)

• Range resolution (x): (c.f. theoretical values, +10% limit & <38m requirement for IS1 and <30m requirement for IS2 - IS7)

20

22

24

26

28

30

32

34

36

38

40

10 15 20 25 30 35 40 45 50

Incidence Angle (Deg)

AP

P R

an

ge

Re

solu

tion

(m

)

IS1

IS2

IS3

IS4

IS5

IS6

IS7

Data



• -12.87±1.44dB (c.f. -12.4dB theoretical value, +5dB limit & <-12dB requirement)

Integrated Sidelobe Ratio (ratio of energy in the sidelobes up to a box 20x by 20y to the energy in the mainlobe(2x by 2y)):

Peak Sidelobe Ratio (ratio of the intensity of the most intense peak outside the main lobe up to 10x by 10y to the energy in the mainlobe):• -19.07±0.94dB (c.f. -21.2dB theoretical value, +5dB

limit & <-20dB requirement)

Spurious Sidelobe Ratio (ratio of the intensity of the most intense peak outside 10x by 10y up to 20x by 20y to the energy in the mainlobe):• -25.98±2.20dB (c.f. <-25dB requirement)



Impulse Response Function Measurements (APM)• IRF parameters have been derived using the ESA transponders in The

Netherlands and the Radarsat transponders in Canada (images 4.8 by 4.8km):

Resolute (HH)Aalsmeer (VH)

Edam (VV)

Zwolle (VH)

Swifterbant (VH) Resolute (HV)Aalsmeer (HV)

Edam (VH)

Edam (HV) Swifterbant (HV)



Example APM Impulse Response Functions

ASAR transponder, i = 25.36°, azimuth resolution = 145.9m, range resolution = 122.1m

Radarsat transponder, i = 44.22°, azimuth resolution = 144.0m (HV) & 143.7m (HH), range resolution = 132.0m (HV) & 133.9m (HH)



Spatial Resolution (3dB width of IRF):• Azimuth resolution (y): 143.9±5.9m (c.f. ~135m theoretical

value & +10% limit)• Range resolution (x): (c.f. theoretical values & +10% limit)

100

110

120

130

140

150

160

170

10 15 20 25 30 35 40 45 50

Incidence Angle (Deg)

AP

M R

an

ge

Re

solu

tion

(m

)

IS1

IS2

IS3

IS4

IS5

IS6

IS7

Data

Note that the measurements are to 1/8 pixel (9.4m) and that APM products are undersampled if the resolution is < 150m.



• -12.37±1.55dB (c.f. -12.4dB theoretical value, +5dB limit & <-12dB requirement)

Integrated Sidelobe Ratio (ratio of energy in the sidelobes up to a box 20x by 20y to the energy in the mainlobe(2x by 2y)):

Peak Sidelobe Ratio (ratio of the intensity of the most intense peak outside the main lobe up to 10x by 10y to the energy in the mainlobe):• -16.62±2.57dB (c.f. -21.2dB theoretical value, +5dB

limit & <-20dB requirement)

Spurious Sidelobe Ratio (ratio of the intensity of the most intense peak outside 10x by 10y up to 20x by 20y to the energy in the mainlobe):• -21.34±2.55dB (c.f. <-25dB requirement)



AP Cross-Polarisation RatioThe ratio of the total power in the IRF of an ASAR transponder in both channels. Examples of IRF in second polarisation:

Aalsmeer (HH)

Edam (HV) Swifterbant (HH)

• Average APP cross-pol ratio (17 measurements): -32.1±4.2dB (c.f. predicted value of < -35dB). Caused by ASAR and/or transponders



AP Channel Co-registration

The mis-registration between the two channels computed from the difference in the location of a point target IRF peak in both channels. As the Resolute transponders give strong IRF’s in all polarisations, they have been used for the co-registration:• APP products (2 measurements): 0.0m & 0.0m

• APM products (2 measurements): 9.4m & 0.0m



Equivalent Number of Looks and Radiometric Resolution

• Mean equivalent number of looks: 1.99±0.05 (c.f. >1.9 requirement, -10% limit)

• Mean radiometric resolution: 2.32±0.03dB (c.f. <2.37dB requirement)

Equivalent number of looks and radiometric resolution are derived using uniform distributed targets. APP co-polarisation

measurements (4 measurements):

APM co-polarisation measurements (16 measurements):

• Mean equivalent number of looks: 56.1±16.8 (c.f. > 30 requirement, -10% limit)

• Mean radiometric resolution: 0.56±0.09dB (c.f. < 0.7dB requirement)



Azimuth Ambiguities

• ASAR Transponders: -27.9±2.6dB

As Doppler frequencies can only be distinguished modulo the PRF, azimuth ambiguities occur within the azimuth antenna pattern sidelobes.

Measurement requires either a very bright point target or a bright point target with a low ambiguity background radar cross-section.

Average APP ambiguity ratio:

The requirement is -25dB while the worst case prediction is ~ -27.9dB.

• ASAR Transponders: -28.7±1.9dB

Average APM ambiguity ratio:

APM Edam (VH)



Localisation Accuracy

• Mean range displacement: -3.8±26.3m• Mean azimuth displacement: -

12.9±60.1m• Mean displacement: 57.8±30.9m

The difference between the measured and predicted positions of the ASAR transponder. The predicted positions are based on image header parameters, the known location of the transponders and their time delay. The ASAR transponders have a small terrain height and hence a small range terrain displacement.

APP measurements:



The localisation accuracy requirement is <900m while the worst case prediction is ~75m in azimuth and between ~125m (IS1) and ~50m (IS7) in range.

APP

0

20

40

60

80

100

120

140

10 15 20 25 30 35 40 45 50

Incidence Angle (deg.)

To

tal D

isp

lace

me

nt (

m)

-150

-100

-50

0

50

100

-100 -50 0 50

Range Displacement (m)

Azi

mu

th D

isp

lace

me

nt (

m)



Preliminary Radiometric CalibrationThe ASAR APP transponders have been used to give a preliminary radiometric calibration for swaths IS2 to IS7. As the v3.03 AP products seem to have been processed with IM nominal chirp powers, corrections have been applied to all AP radar cross-section measurements.

4

6

8

10

12

14

16

15 20 25 30 35 40 45

Incidence Angle (deg)

AP

M C

hirp

Pow

er (

dB)

IS1

IS2

IS3

IS4

IS6

IS7

IS5

Data



-1.5

-1.0

-0.5

0.0

0.5

1.0

15 20 25 30 35 40 45 50


Cor

rect

ed A

PP

Rel

ativ

e R

CS

(dB

)

NLNL SatNL

• APP K for swaths IS2 to IS4 is: 57.05±0.51dB• APP K for swaths IS5 to IS7 is: 60.59±0.47dB (based on non-

saturated ASAR transponders)• Different K values due to different product scaling factors



The derived APM calibration K is: 69.47±0.50dB (based on ASAR transponders with acceptable ISLR’s)

Further measurements required before definitive calibration constants can be derived

-1.0

-0.5

0.0

0.5

1.0

15 20 25 30 35 40 45


Cor

rect

ed A

PM

Rel

ativ

e R

CS

(dB

)



Noise Equivalent Radar Cross-sectionNESigma0 estimated using low radar cross-section regions (ocean, open water or lakes). This gives an upper limit to NESigma0. HV APP polarisation measurements:

-30

-28

-26

-24

-22

-20

-18

14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46Incidence Angle (degs)

AP

P H

V N

ES

igm

a0

(d

B)



Summary• 10 v3.03 APP and APM IS2 to IS7 products analysed• No format problems identified but many of the APP products had

no Chirp Parameter ADSR (hence no annotated Chirp Powers)• No visualisation problems found with any post Orbit 3661 v3.03

products• APP azimuth & range resolutions, ISLR, PSLR & SSLR acceptable.

Some range under-sampling• APM Azimuth & range resolutions and ISLR acceptable. PSLR &

SSLR outside expected range due to APM under-sampling in both azimuth and at almost all ground ranges

• Acceptable APP cross-polarisation ratio• Sub-pixel AP channel co-registration



• APP & APM equivalent number of looks and radiometric resolution acceptable

• APP & APM azimuth ambiguities acceptable• APP Localisation accuracy good • Preliminary calibration constants derived - more work

required before definitive K values can be calculated. • Noise equivalent radar cross-sections lower than predicted

NESigma0

ASAR APP & APM Image Quality Peter Meadows & Trish Wright

Documents