Remote Sensing Applications and Decision Support

Realization of strong backscattering homogeneous regions with known backscattering coefficient in synthetic aperture radar images

[+] Author Affiliations
Xin Lin, Kaizhi Wang, Junfeng Wang, Xingzhao Liu

Shanghai Jiao Tong University, Department of Electronic Engineering, 800 Dongchuan Road, Shanghai 200240, China

J. Appl. Remote Sens. 11(1), 016018 (Feb 01, 2017). doi:10.1117/1.JRS.11.016018
History: Received September 3, 2016; Accepted January 10, 2017
Text Size: A A A

Open Access Open Access

Abstract.  The strong backscattering homogeneous region, i.e., a uniform region with a high and constant backscattering coefficient, is important for synthetic aperture radar (SAR) image quality assessment and SAR radiometric calibration, which, however, is difficult to realize in practice with a known backscattering coefficient. We realize a strong backscattering homogeneous region with a known backscattering coefficient in SAR images by utilizing designed metal grids. First, we propose a manmade grid-structure target and realize it with aluminum in practice, which is named the metal grid. Then, the backscattering coefficient of the designed metal grid is simulated in the computer simulation technology (CST) microwave studio and measured by a radar cross-section (RCS) measurement instrument in a microwave anechoic chamber. Both CST simulation results and RCS measurement results confirm the strong backscattering property of the designed target. In addition, by utilizing the designed target, we realize a test field consisting of several strong backscattering homogeneous regions with different sizes at Shanghai Jiao Tong University, Shanghai, China. The spaceborne experiments have been carried out by the TerraSAR-X sensor over the test field in two flight campaigns in X-band with VV polarization. Experimental results demonstrate the strong backscattering property and homogeneity of the realized regions.

Figures in this Article

During the last two decades, the application of synthetic aperture radar (SAR) images has been greatly extended, owing to the successful launches of high-resolution SAR satellites and the advances in SAR technology.1,2 SAR image quality assessment, which aims to evaluate the imaging performance of the SAR system, is usually done by measuring the impulse response function based on an isolated point target positioned in the scene. However, the complexity of manmade and natural targets in SAR images is usually significant, so they cannot be characterized using the simplistic point target. In recent years, to assess the imaging performance for real scenarios accurately, some research has introduced the modulation transfer function (MTF) of electro optical (EO) sensors into SAR systems, which has been widely used in optical imaging systems and proven as an effective approach for assessing image quality.35 For SAR systems, the measurement of the MTF is based on a test field consisting of homogeneous regions6 with strong backscattering property, i.e., uniform regions with high and constant backscattering coefficients. Reference 3 studied the MTF-based image quality assessment method for SAR systems, where the involved sine-wave test field is difficult to realize in practice due to the strict backscattering requirement. The authors in Ref. 4 constructed a test field by utilizing the surface roughness, where the strong backscattering region is realized by coarse gravels. However, the homogeneity and backscattering property of the test field are highly influenced by the environmental conditions (weather and humidity). In our previous work,5 we proposed an MTF-based image quality assessment method for SAR image interpretation. However, this work mainly focused on the MTF calculation and the utilization of the MTF curve, not on the realization of strong backscattering homogeneous regions.

Furthermore, homogeneous regions are of great importance in the calibration activities of the SAR satellites in orbit, which are used to measure some radiometric indicators such as the radiometric resolution and the relative radiometric accuracy. In general, SAR radiometric calibration can be completed based on natural homogeneous regions, e.g., lawns, forests, seas, and rainforests, and it assumes that the selected region is uniform with a known average backscattering coefficient.7,8 In Ref. 8, these radiometric indicators are measured based on the Amazon rainforests. The flatness and isotropy of the Amazon rainforest allowed an approximate homogeneous characterization, which makes this region a reference site in the in-orbit radiometric calibration for a number of SAR satellites, e.g., RADARSAT-1 and RADARSAT-2. However, these natural targets usually appear as weak backscattering regions in SAR images. Under some conditions of high noise level, they may be obscured by noise and no longer be suitable for SAR radiometric calibration. In addition, due to their natural properties, the backscattering coefficients of these natural targets are uncontrollable, and the homogeneity is also influenced by environment factors, e.g., the homogeneity of the Amazon rainforest suffers from the moisture variation and temperature variation throughout the year. These may incur additional errors in SAR radiometric calibration.

In this paper, we realize a strong backscattering homogeneous region with a known backscattering coefficient in SAR images by utilizing the designed metal grids. First, we design a manmade grid-structure target and realize it with aluminum in practice, which is referred to as the metal grid in this paper. Then, to verify the strong backscattering property of the designed metal grid, its backscattering coefficient is both simulated in the computer simulation technology (CST) microwave studio and measured by a radar cross-section (RCS) measurement instrument in a microwave anechoic chamber. Both the CST simulation results and the RCS measurement results confirm its strong backscattering property. After that, we realize a test field consisting of several strong backscattering homogeneous regions with different sizes at Shanghai Jiao Tong University, Shanghai, China. The spaceborne experiments have been carried out by the TerraSAR-X sensor in X-band with VV polarization (X-VV) over the test field in two flight campaigns, and the corresponding experimental results also confirm the strong backscattering property and the homogeneity of the realized region.

In this paper, Sec. 2 briefly describes the measurement theory of the backscattering coefficient. Section 3 introduces the designed metal grid. Sections 4 and 5 present the CST simulation results and the RCS measurement results of the designed metal grid, respectively. The spaceborne experimental results in Sec. 6 confirm the benefits of the proposed target. Section 7 gives some discussions, and Sec. 8 concludes this work.

The aim of our work is to realize strong backscattering homogenous regions in SAR images. Therefore, to analyze the scattering property of the designed grid-structure target, the backscattering coefficient of the proposed strong backscattering homogeneous region needs to be measured from a focused single-look slant-range complex (SSC) SAR image. In general, the backscattering coefficient of a homogeneous region depends on the radar parameters (transmitting frequency f and polarization mode p), the measurand characteristics m (structure and material), and the geometry parameter (incidence angle θ).9,10Figure 1 shows the definition of the incidence angle θ, which is defined as the angle between the line of sight of the radar sensor and the normal to the Earth surface. Here we assume that the earth surface is a flat surface for simplicity. Thus, the backscattering coefficient of a homogeneous region is determined by a number of parameters and can be written in the following way: σ0(f,p,m,θ).

Graphic Jump Location
Fig. 1
F1 :

Definition of the incidence angle θ, which is defined as the angle between the line of sight of the radar sensor and the normal to the Earth surface.

The SAR intensity image I(x,y) can be obtained from the SSC SAR image S(x,y) after a square-law detection,11 i.e., I(x,y)=|S(x,y)|2. Neglecting noise, for a homogeneous region H, the integral of the image intensity I(x,y) is represented as Display Formula

Es=AHI(x,y)dxdy,(1)
Display Formula
=KAHσ0,(2)
where K is the absolute calibration coefficient, AH is the area of H, and σ0 is the backscattering coefficient of H. After radiometric calibration, the pixel intensity is proportional to the backscattering coefficient (for homogeneous regions) or the radar cross section (for point targets).11 Therefore, the backscattering coefficient of a homogeneous region H can be calculated by Display Formula
σ0=EsKAH,(3)
Display Formula
=AHI(x,y)dxdyKAH.(4)
Thus, after obtaining the integral Es of the image intensity I(x,y) over a homogeneous region H, one can measure its backscattering coefficient σ0 by Eqs. (3) and (4). In the following parts, we will first introduce the designed metal grid and then analyze its backscattering property and homogeneity based on the aforementioned definitions.

Based on the analysis over a set of SAR images from different SAR sensors and imaging modes, e.g., stripmap mode, ScanSAR mode, and spotlight mode, we find an interesting phenomenon that the grid structure has always shown a strong backscattering property in SAR images.12Figure 2 gives the TerraSAR-X 3-m image (stripmap mode) and the corresponding optical images (YaoGan-5 optical image and Google Earth image) of the third runway construction process in Tokyo Haneda Airport, Japan, where the grid-structure targets used for the runway construction in Fig. 2(c) appear as a strong backscattering homogeneous region in the TerraSAR-X image.

Graphic Jump Location
Fig. 2
F2 :

TerraSAR-X 3-m image and the corresponding optical images of the runway construction in Tokyo Haneda Airport: (a) TerraSAR-X 3-m image, (b) YaoGan-5 optical image, and (c) Google Earth image.

The aim of our work is to design a manmade target with a strong backscattering property and a symmetrical structure that can be easily extended to a large area. In this paper, we design a manmade target with the grid structure to realize a strong backscattering homogeneous region in SAR images, which is named the metal grid. The detailed structure of the designed target is shown as Fig. 3. In Fig. 3, the designed target consists of a number of small square grids with the same size, where the size of the designed grid-structure target is L×L×h and the size of each small square grid is L0×L0×h. In addition, the gray color denotes the material of the designed target, which is typically made of metal for better backscattering property.

Graphic Jump Location
Fig. 3
F3 :

The schematic diagram of the designed grid-structure target (L×L×h), which shows a strong backscattering property in SAR images. Each small square grid is L0×L0×h.

Figure 4 gives an example of the designed grid-structure target (i.e., metal grid), which is made of aluminum. In Fig. 4, the size of the designed metal grid is 2×2×0.09  m, and each small square grid is 15×15×9  cm. The metal grid is assembled by the aluminum strips with designed slots at fixed intervals, and the strips are combined with each other by these slots. These aluminum strips are purchased from the building materials market with the mm-level fabrication precision, and the cost of the metal grid target in Fig. 4 is about 5 dollars. Since the needed material of the designed metal grid can be easily obtained from building material markets at low cost, it provides a feasible way to realize the strong backscattering homogeneous region in practice.

Graphic Jump Location
Fig. 4
F4 :

An example of the designed metal grid (2×2×0.09  m), which is utilized to realize a strong backscattering homogeneous region in SAR images. Each small square grid is 15×15×9  cm.

In general, the backscattering coefficient σ0 of the designed metal grid can be estimated in three different ways: simulated by the CST microwave studio, measured by an RCS measurement instrument in a microwave anechoic chamber, and measured experimentally from SAR image. In this paper, we consider all three ways to verify the strong backscattering property of the designed metal grid.

In this section, to verify the strong backscattering property of the designed metal grid, its electromagnetic scattering property is simulated in the CST microwave studio,13 which is widely used for electromagnetic design and analysis. Figure 5 gives the geometric model of the designed metal grid (made of aluminum) used for the electromagnetic scattering simulation in the CST microwave studio. The size of the metal grid is 2×2×0.09  m, and each small square grid is 15×15×9  cm.

Graphic Jump Location
Fig. 5
F5 :

The geometric model of the designed metal grid (2×2×0.09  m) used for the electromagnetic scattering simulation in the CST microwave studio. Each small square grid is 15×15×9  cm.

The electromagnetic scattering property of the metal grid is simulated at θ0=30  deg and θ0=45  deg incidence angles with 9.6-GHz transmitting frequency, where θ0 is defined as the angle between the incident electromagnetic wave and the normal to the object surface. In addition, the aspect angle (viewing perspective projected on a horizontal plane) is selected as the line of sight along one target axis. Here we consider 9.6 GHz, for it is close to the center frequency of the TerraSAR-X sensor. The electromagnetic scattering simulation results of the proposed metal grid at 30-deg and 45-deg incidence angles are shown in Fig. 6. As can be seen, for the 30-deg incidence angle in Fig. 6(a), two strong electromagnetic scattering directions of the metal grid exist at 30 deg and 150 deg. In addition, it can be found that the maximum RCS exists at 30 deg as 47.2  dBm2 (shown as the blue line) with the side lobe level as 21.4  dB. This indicates that the strongest scattering direction of the designed metal grid is the same as the incidence angle, which verifies its strong backscattering property. Similarly, for the 45-deg incidence angle in Fig. 6(b), it can be found that the maximum RCS exists at 45 deg (the same as its incidence angle) as 45.8  dBm2, shown as the blue line, and the side lobe level is 18.6  dB. Therefore, from the CST simulation results, we find that the maximum scattering direction (with the maximum RCS) of the designed target is the same as the incidence angle θ0, which verifies the strong backscattering property of the designed metal grid.

Graphic Jump Location
Fig. 6
F6 :

The CST simulation results of the designed metal grid (2×2×0.09  m) at different incidence angles with 9.6-GHz transmitting frequency: (a) 30-deg incidence angle and (b) 45-deg incidence angle.

In this section, we further investigate the backscattering property of the designed metal grid with regard to the frequency f and the incidence angle θ. The RCS is measured by the OP240-AB RCS measurement instrument (MARCH Corporation) in a microwave anechoic chamber in Shanghai, China, as shown in Fig. 7. The size of the metal grid is 2×2×0.09  m, and each small square grid is 15×15×9  cm. The designed metal grid is hung on the pillar on a fully automated 360 deg rotatable platform placed in the middle of the microwave anechoic chamber (see Fig. 7). Both the platform and the pillar are masked with radar-absorbing materials to ensure the accurate RCS measurement of the metal grid. The RCS of the metal grid is measured from 0-deg to 90-deg incidence angle with a 0.2-deg step. Furthermore, the monostatic configuration of the RCS measurement instrument measures the reflected signal from 8 to 12 GHz with a 2-MHz step, in the X-band with VV polarization. Here we consider 8-12 GHz and VV polarization to correspond to the transmitting frequency (9.65 GHz) and the polarization mode (VV) of TerraSAR-X sensor in the spaceborne experiments. All the returned signals of the metal grid are calibrated according to the RCS of a calibration sphere, which was measured before our experiment.

Graphic Jump Location
Fig. 7
F7 :

The RCS measurement setup for the designed metal grid (2×2×0.09  m) in a microwave anechoic chamber.

The RCS measurement results of the designed metal grid at 30-deg, 45-deg, and 60-deg incidence angles are shown in Fig. 8. From Fig. 8, we find that for all incidence angles, the maximum RCS values of the designed metal grid are larger than 21.3  dBm2 and appear around 9.6 to 10 GHz, which is very close to the center transmitting frequency (9.65 GHz) of the TerraSAR-X sensor. In addition, it can be found that the designed target always has large RCS values at most X-band frequencies for the 30-deg, 45-deg, and 60-deg incidence angles. From the measurement results, we can see that the RCS of the designed target varies in the range of 10 to 25  dBm2, at 30 deg to 60-deg incidence angle in 9 to 10.5 GHz frequency with VV polarization. Thus, it is feasible to utilize the designed metal gird to realize a strong backscattering homogeneous region in SAR images. In addition, due to the symmetrical grid-structure, one may expect that the geometry and the size of the designed metal grid will have little limits on the frequency selection. In future work, the backscattering property of the designed metal grid in other bands (L/C/S band) and polarization modes (HH/HV/VH) will be further studied.

Graphic Jump Location
Fig. 8
F8 :

The RCS measurement results of the metal grid at the fixed incidence angles from 8 to 12 GHz frequency in the X-band with VV polarization. The incidence angles are (a) 30 deg, (b) 45 deg, and (c) 60 deg.

In this section, some spaceborne experimental results obtained by TerraSAR-X are given to demonstrate the effectiveness of realizing a strong backscattering homogeneous region in SAR images using our designed metal grids.

Test Field

As mentioned in Sec. 3, the strong backscattering homogeneous region can be realized by the designed metal grids in Fig. 4 at low cost. The backscattering coefficient (at a fixed incidence angle and a fixed transmitting frequency, i.e., 9.65 GHz) of the metal grid can be measured in a microwave anechoic chamber, which is shown in Fig. 8. Then, we realize a 50×48  m test field constructed on a 300×200  m vacant area (against the background of bare soil) at Shanghai Jiao Tong University, Shanghai, China, in October 2015, as shown in Fig. 9. The designed test field consists of several strong backscattering regions of different sizes, and the sketch map is shown as Fig. 9(b). In Fig. 9(b), the widths of the strong backscattering regions (denoted as the white bars) are 0.5, 0.5, 1, 2, 4, 6, 8, and 2 m from bottom to top with 50 m length, and the black bars denote the bare soil. Thus, the size of the largest region is 50×8  m, resulting in a 400  m2 strong backscattering homogeneous region. In addition, the designed test field can also be used for measuring the MTF of SAR systems for SAR image quality assessment.5 The total cost of the test field in Fig. 9 is about 2000 dollars, including 1500 dollars for the metal grid targets (1200  m2 in total) and 500 dollars for four workers to assemble the grid targets and construct the test field in 3 days.

Graphic Jump Location
Fig. 9
F9 :

A test field constructed at Shanghai Jiao Tong University, Shanghai, China, where the metal grids are placed against the background of bare soil: (a) and (b) the sketch maps of the test field, (c)–(e) photos of the test field, and (f) the strong backscattering regions realized by the designed metal grids.

Spaceborne Experiments

In this section, the German satellite-borne SAR sensor TerraSAR-X is used in two different flight campaigns for the acquisition of the SAR image data. This SAR sensor has been in use for various applications over a period of 9 years, e.g., maritime surveillance, disaster monitoring, and natural resource management.14,15 The experimental data are collected by TerraSAR-X during two flight campaigns: the first on November 25, 2015, and the second on November 29, 2015, both over the test field in Fig. 9 realized at Shanghai Jiao Tong University in China. These two flight campaigns are carried out in different imaging modes (spotlight and stripmap modes) but with the same VV polarization in X-band. The incidence angles of two flight campaigns were 40.14 deg and 41.73 deg. Table 1 lists the main TerraSAR-X system parameters used in these two flight campaigns, which is provided by the TerraSAR-X product data. The footprints of these SAR data acquisitions are shown in Fig. 10.

Table Grahic Jump Location
Table 1TerraSAR-X system parameters in two experiments.
Graphic Jump Location
Fig. 10
F10 :

Footprints of two SAR data acquisitions over Shanghai Jiao Tong University, Shanghai, China.

Experimental Results

Figure 11 shows the TerraSAR 2-look intensity images of the realized test field, consisting of several strong backscattering regions with different widths, which are obtained after focusing and calibration. Figure 11(a) is acquired in the spotlight mode during the first flight campaign, and Fig. 11(b) gives the details of the test field in Fig. 11(a). Figure 11(d) is acquired in the stripmap mode during the second flight campaign, and Fig. 11(e) gives the details of the test field in Fig. 11(d). As can be seen, the realized strong backscattering regions can be clearly distinguished and appear as uniform white regions. Specifically, it can be found that the white bars of 0.5 and 1 m width overlap each other and mask the background (bare soil) in Fig. 11(b) due to the resolution limitation of the SAR system. Similarly, the white bars of 0.5, 1, 2, and 4 m width in Fig. 11(e) cannot be distinguished clearly and are shown as one white region (highlighted by the yellow frame) due to the limited system resolution. The detailed explanation of this phenomenon can be found in our previous work.5

Graphic Jump Location
Fig. 11
F11 :

TerraSAR-X images of the test field at Shanghai Jiao Tong University: (a) the SAR image acquired in the first experiment (spotlight mode), (b) the zoomed-in image of (a), (c) the backscattering coefficients of the 8 m width white bar (50×8  m) in (b), (d) the SAR image acquired in the second experiment (stripmap mode), (e) the zoomed-in image of (d), and (f) the backscattering coefficients of the 8 m width white bar (50×8  m) in (e).

Then, we can measure the backscattering coefficient σ0 of the strong backscattering homogeneous regions from the focused SAR intensity images by Eqs. (3) and (4). Figures 11(c) and 11(f) show the backscattering coefficients of the largest white bar with 8 m width (50×8  m) in Figs. 11(b) and 11(e), respectively. As can be seen, for the first experiment (spotlight mode) in Fig. 11(c), the mean backscattering coefficient σ0 of the strong backscattering region is 7.18 dB with the standard deviation as 4.81 dB. For the second experiment (stripmap mode) in Fig. 11(f), the mean backscattering coefficient σ0 is 6.74 dB with the standard deviation as 3.91 dB. In addition, the backscattering coefficient of the background (the bare soil) is about 0.19 dB. By analyzing the experiment results, we find that the designed metal grid has a high backscattering coefficient value of about 7 dB at the fixed 40 deg–41 deg incidence angle in X-VV.

Furthermore, the coefficient of variance (CV) is introduced here to evaluate the homogeneity of the strong backscattering regions realized by the designed metal grids. In general, the CV is calculated to decide the homogeneity or inhomogeneity of a region.16,17 The CV is defined as Display Formula

CV=σμ,(5)
where σ and μ are the standard deviation and the mean value of the estimated region, respectively. Generally, the CV is a constant in statistics for homogeneous scenes and only depends on the number of looks of the SAR image.16 For an L-look SAR intensity image, the CV of a homogeneous region in statistics is given by Ref. 17Display Formula
CV=σμ=1L.(6)
Considering the experimental results in Fig. 11 (2-look TerraSAR-X intensity images), the CV values of the realized strong backscattering regions in Figs. 11(b) and 11(e) are calculated to be 0.67 and 0.58, respectively, which are both close to the expected CV value of the homogeneous region in a 2-look intensity image as CV=12=0.707. This means that the intensity fluctuations in the realized regions are mainly caused by the speckle noise, which is an inherent characteristic of the SAR image. Therefore, the CV values verify the homogeneity of the strong backscattering regions realized by the proposed target.

As previously mentioned, the backscattering coefficient of the designed metal grid at a fixed incidence angle and a fixed frequency can be measured in the microwave anechoic chamber. With the known backscattering coefficient, the strong backscattering homogeneous region could be useful for many SAR applications, e.g., SAR image quality assessment and SAR radiometric calibration. For the SAR image quality assessment, based on the realized strong backscattering homogeneous regions, one can directly measure the MTF of SAR systems, which provides a feasible way to utilize the general assessment methods of EO sensors for the SAR image quality assessment.5

In addition, with the known backscattering coefficient, the realized strong backscattering homogeneous region can also be utilized for SAR radiometric calibration. Generally, SAR radiometric calibration can be completed based on natural homogeneous regions, e.g., lawns, forests, seas, and rainforests. However, the homogeneity of these natural targets suffers from the moisture variation and the temperature variation, which may incur unpredictable errors in SAR radiometric calibration. In constrast, the strong backscattering homogeneous region realized by the proposed metal grid has better temporal stability, flatness, and isotropy because of its manmade characteristic. In addition, different than the uncontrollable backscattering coefficient of these natural targets, the proposed manmade target has a constant backscattering coefficient and its backscattering coefficient can be premeasured in a microwave anechoic chamber, as shown in Sec. 5, ensuring a more accurate SAR radiometric calibration. Moreover, under some conditions of high noise level, these conventional targets with weak backscattering property may be obscured by the noise and no longer suitable for SAR radiometric calibration, while the proposed target with strong backscattering property is robust against high noise level for SAR radiometric calibration.

One may also consider utilizing a rotated metal grid as an alternative for realizing strong backscattering homogeneous regions, which may bring two main benefits. First, compared with the metal grid placed along azimuth (target signal is the sum of dihedrals), the rotated metal grid may have a stronger backscattering property since its target signal is the sum of trihedrals. Second, comparing to the metal grid with the specific aspect angle selected as a line of sight along one target axis, the potential error-prone grid positioning along the azimuth direction will no longer be necessary. However, for some specific applications, e.g., measuring the MTF of SAR system, the rotated metal grid is unsuitable due to the alignment requirement of the target axis and the azimuth direction.5

In this paper, with the proposed metal grids, we realize a strong backscattering homogeneous region with a known backscattering coefficient in SAR images. Both the CST simulation results and the RCS measurement results confirm the strong backscattering property of the designed metal grid. In addition, a test field consisting of several strong backscattering homogeneous regions with different sizes was constructed at Shanghai Jiao Tong University, Shanghai, China. The spaceborne experiments were carried out by the TerraSAR-X sensor over the test field in two flight campaigns. The experimental results verify both the strong backscattering property and the homogeneity of the realized region. With the known backscattering coefficient, the realized strong backscattering homogeneous region could be useful for many SAR applications, such as SAR image quality assessment and radiometric calibration.

This work was supported by National Natural Science Foundation of China Grant No: 61132005.

Moreira  A.  et al., “A tutorial on synthetic aperture radar,” IEEE Geosci. Remote Sens. Mag.. 1, (1 ), 6 –43 (2013).CrossRef
Chen  J., , Zhang  B., and Wang  C., “Backscattering feature analysis and recognition of civilian aircraft in TerraSAR-X images,” IEEE Geosci. Remote Sens. Lett.. 12, (4 ), 796 –800 (2015).CrossRef
Lin  X.  et al., “A new MTF-based image quality assessment for high-resolution SAR sensors,” in  Int. Geoscience and Remote Sensing Symp. (IGARSS ‘13) , pp. 1305 –1308 (2013).CrossRef
Schulz  K.  et al., “A new concept for SAR image quality assessment realizing bar patterns in SAR images,” in  7th European Conf. on Synthetic Aperture Radar (EUSAR)  (2008).
Lin  X.  et al., “Image quality assessment based on modulation transfer function for synthetic aperture radar systems,” J. Appl. Remote Sens.. 10, (2 ), 026023  (2016).CrossRef
Martinez  A., and March  J., “SAR image quality assessment,” Revista de Teledeteccion. 2, , 12 –18 (1993).
Freeman  A., “SAR calibration: an overview,” IEEE Trans. Geosci. Remote Sens.. 30, (6 ), 1107 –1121 (1992). 0196-2892 CrossRef
Cote  S.  et al., “SAR image quality and calibration operations for the RADARSAT satellites at the Canadian space agency,” in  Int. Radar Conf.: Surveillance for a Safer World  (2009).
Palubinskas  G., , Runge  H., and Reinartz  P., “Radar signatures of road vehicles: airborne SAR experiments,” Proc. SPIE. 5980, , 598006  (2005). 0277-786X CrossRef
Palubinskas  G., and Runge  H., “Radar signatures of a passenger car,” IEEE Geosci. Remote Sens. Lett.. 4, (4 ), 644 –648 (2007).CrossRef
Döring  B. J., and Schwerdt  M., “The radiometric measurement quantity for SAR images,” IEEE Trans. Geosci. Remote Sens.. 51, (12 ), 5307 –5314 (2013). 0196-2892 CrossRef
Lin  X.  et al., “A novel concept of plane grid resolution for high-resolution SAR imaging systems,” in  Int. Geoscience and Remote Sensing Symp. (IGARSS) , pp. 358 –361 (2014).CrossRef
CST–Computer Simulation Technology, “CST studio suite 2016,” https://www.cst.com/Content/Articles/article909/CST-STUDIO-SUITE-2016.pdf
Mittermayer  J.  et al., “TerraSAR-X system performance characterization and verification,” IEEE Trans. Geosci. Remote Sens.. 48, (2 ), 660 –676 (2010). 0196-2892 CrossRef
Pitz  W., and Miller  D., “The TerraSAR-X satellite,” IEEE Trans. Geosci. Remote Sens.. 48, (2 ), 615 –622 (2010). 0196-2892 CrossRef
Oliver  C., and Quegan  S., Understanding Synthetic Aperture Radar Images. ,  Artech House ,  Boston, Massachusetts  (1998).
Beauchemin  M., , Thomson  K., and Edwards  G., “The ratio of the arithmetic to the geometric mean: a first-order statistical test for multilook SAR image homogeneity,” IEEE Trans. Geosci. Remote Sens.. 34, (2 ), 604 –606 (1996). 0196-2892 CrossRef
© The Authors. Published by SPIE under a Creative Commons Attribution 3.0 Unported License. Distribution or reproduction of this work in whole or in part requires full attribution of the original publication, including its DOI.

Citation

Xin Lin ; Kaizhi Wang ; Junfeng Wang and Xingzhao Liu
"Realization of strong backscattering homogeneous regions with known backscattering coefficient in synthetic aperture radar images", J. Appl. Remote Sens. 11(1), 016018 (Feb 01, 2017). ; http://dx.doi.org/10.1117/1.JRS.11.016018


Figures

Graphic Jump Location
Fig. 1
F1 :

Definition of the incidence angle θ, which is defined as the angle between the line of sight of the radar sensor and the normal to the Earth surface.

Graphic Jump Location
Fig. 2
F2 :

TerraSAR-X 3-m image and the corresponding optical images of the runway construction in Tokyo Haneda Airport: (a) TerraSAR-X 3-m image, (b) YaoGan-5 optical image, and (c) Google Earth image.

Graphic Jump Location
Fig. 3
F3 :

The schematic diagram of the designed grid-structure target (L×L×h), which shows a strong backscattering property in SAR images. Each small square grid is L0×L0×h.

Graphic Jump Location
Fig. 4
F4 :

An example of the designed metal grid (2×2×0.09  m), which is utilized to realize a strong backscattering homogeneous region in SAR images. Each small square grid is 15×15×9  cm.

Graphic Jump Location
Fig. 5
F5 :

The geometric model of the designed metal grid (2×2×0.09  m) used for the electromagnetic scattering simulation in the CST microwave studio. Each small square grid is 15×15×9  cm.

Graphic Jump Location
Fig. 6
F6 :

The CST simulation results of the designed metal grid (2×2×0.09  m) at different incidence angles with 9.6-GHz transmitting frequency: (a) 30-deg incidence angle and (b) 45-deg incidence angle.

Graphic Jump Location
Fig. 7
F7 :

The RCS measurement setup for the designed metal grid (2×2×0.09  m) in a microwave anechoic chamber.

Graphic Jump Location
Fig. 8
F8 :

The RCS measurement results of the metal grid at the fixed incidence angles from 8 to 12 GHz frequency in the X-band with VV polarization. The incidence angles are (a) 30 deg, (b) 45 deg, and (c) 60 deg.

Graphic Jump Location
Fig. 9
F9 :

A test field constructed at Shanghai Jiao Tong University, Shanghai, China, where the metal grids are placed against the background of bare soil: (a) and (b) the sketch maps of the test field, (c)–(e) photos of the test field, and (f) the strong backscattering regions realized by the designed metal grids.

Graphic Jump Location
Fig. 10
F10 :

Footprints of two SAR data acquisitions over Shanghai Jiao Tong University, Shanghai, China.

Graphic Jump Location
Fig. 11
F11 :

TerraSAR-X images of the test field at Shanghai Jiao Tong University: (a) the SAR image acquired in the first experiment (spotlight mode), (b) the zoomed-in image of (a), (c) the backscattering coefficients of the 8 m width white bar (50×8  m) in (b), (d) the SAR image acquired in the second experiment (stripmap mode), (e) the zoomed-in image of (d), and (f) the backscattering coefficients of the 8 m width white bar (50×8  m) in (e).

Tables

Table Grahic Jump Location
Table 1TerraSAR-X system parameters in two experiments.

References

Moreira  A.  et al., “A tutorial on synthetic aperture radar,” IEEE Geosci. Remote Sens. Mag.. 1, (1 ), 6 –43 (2013).CrossRef
Chen  J., , Zhang  B., and Wang  C., “Backscattering feature analysis and recognition of civilian aircraft in TerraSAR-X images,” IEEE Geosci. Remote Sens. Lett.. 12, (4 ), 796 –800 (2015).CrossRef
Lin  X.  et al., “A new MTF-based image quality assessment for high-resolution SAR sensors,” in  Int. Geoscience and Remote Sensing Symp. (IGARSS ‘13) , pp. 1305 –1308 (2013).CrossRef
Schulz  K.  et al., “A new concept for SAR image quality assessment realizing bar patterns in SAR images,” in  7th European Conf. on Synthetic Aperture Radar (EUSAR)  (2008).
Lin  X.  et al., “Image quality assessment based on modulation transfer function for synthetic aperture radar systems,” J. Appl. Remote Sens.. 10, (2 ), 026023  (2016).CrossRef
Martinez  A., and March  J., “SAR image quality assessment,” Revista de Teledeteccion. 2, , 12 –18 (1993).
Freeman  A., “SAR calibration: an overview,” IEEE Trans. Geosci. Remote Sens.. 30, (6 ), 1107 –1121 (1992). 0196-2892 CrossRef
Cote  S.  et al., “SAR image quality and calibration operations for the RADARSAT satellites at the Canadian space agency,” in  Int. Radar Conf.: Surveillance for a Safer World  (2009).
Palubinskas  G., , Runge  H., and Reinartz  P., “Radar signatures of road vehicles: airborne SAR experiments,” Proc. SPIE. 5980, , 598006  (2005). 0277-786X CrossRef
Palubinskas  G., and Runge  H., “Radar signatures of a passenger car,” IEEE Geosci. Remote Sens. Lett.. 4, (4 ), 644 –648 (2007).CrossRef
Döring  B. J., and Schwerdt  M., “The radiometric measurement quantity for SAR images,” IEEE Trans. Geosci. Remote Sens.. 51, (12 ), 5307 –5314 (2013). 0196-2892 CrossRef
Lin  X.  et al., “A novel concept of plane grid resolution for high-resolution SAR imaging systems,” in  Int. Geoscience and Remote Sensing Symp. (IGARSS) , pp. 358 –361 (2014).CrossRef
CST–Computer Simulation Technology, “CST studio suite 2016,” https://www.cst.com/Content/Articles/article909/CST-STUDIO-SUITE-2016.pdf
Mittermayer  J.  et al., “TerraSAR-X system performance characterization and verification,” IEEE Trans. Geosci. Remote Sens.. 48, (2 ), 660 –676 (2010). 0196-2892 CrossRef
Pitz  W., and Miller  D., “The TerraSAR-X satellite,” IEEE Trans. Geosci. Remote Sens.. 48, (2 ), 615 –622 (2010). 0196-2892 CrossRef
Oliver  C., and Quegan  S., Understanding Synthetic Aperture Radar Images. ,  Artech House ,  Boston, Massachusetts  (1998).
Beauchemin  M., , Thomson  K., and Edwards  G., “The ratio of the arithmetic to the geometric mean: a first-order statistical test for multilook SAR image homogeneity,” IEEE Trans. Geosci. Remote Sens.. 34, (2 ), 604 –606 (1996). 0196-2892 CrossRef

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Related Book Chapters

Topic Collections

Advertisement
  • Don't have an account?
  • Subscribe to the SPIE Digital Library
  • Create a FREE account to sign up for Digital Library content alerts and gain access to institutional subscriptions remotely.
Access This Article
Sign in or Create a personal account to Buy this article ($20 for members, $25 for non-members).
Access This Proceeding
Sign in or Create a personal account to Buy this article ($15 for members, $18 for non-members).
Access This Chapter

Access to SPIE eBooks is limited to subscribing institutions and is not available as part of a personal subscription. Print or electronic versions of individual SPIE books may be purchased via SPIE.org.