Improved GNSS-Based Bistatic SAR Using Multi-Satellites Fusion: Analysis and Experimental Demonstration

Abstract

The Global Navigation Satellite System (GNSS)-based Bistatic Synthetic Aperture Radar (SAR) is getting more and more attention in remote sensing for its all-weather and real-time global observation capability. Its low range resolution results from the narrow signal bandwidth limits in its development. The configuration difference caused by the illumination angle and movement direction of the different satellites makes it possible to improve resolution by multi-satellite fusion. However, this also introduces new problems with the resolution-enhancing efficiency and increased computation brought about by the fusion. In this paper, we aim at effectively improving the resolution of the multi-satellite fusion system. To this purpose, firstly, the Point Spread Function (PSF) of the multi-satellite fusion system is analyzed, and focusing on the relationship between the fusion resolution and the geometric configuration and the number of satellites. Numerical simulation results show that, compared with multi-satellite fusion, dual-satellite fusion is a combination with higher resolution enhancement efficiency. Secondly, a method for dual-satellite fusion imaging based on optimized satellite selection is proposed. With the greedy algorithm, the selection is divided into two steps: in the first step, according to geometry configuration, the single-satellite with the optimal 2-D resolution is selected as the reference satellite; in the second step, the angles between the azimuthal vector of the reference satellite and the azimuthal vector of the other satellites were calculated by the traversal method, the satellite corresponding to the intersection angle which is closest to 90° is selected as the auxiliary satellite. The fused image was obtained by non-coherent addition of the images generated by the reference satellite and the auxiliary satellite, respectively. Finally, the GPS L1 real orbit multi-target simulation and experimental validation were conducted, respectively. The simulation results show that the 2-D resolution of the images produced by our proposed method is globally optimal 15 times and suboptimal 8 times out of 24 data sets. The experimental results show that the 2-D resolution of our proposed method is optimal in the scene, and the area of the resolution unit is reduced by 70.1% compared to the single-satellite’s images. In the experiment, there are three navigation satellites for imaging, the time taken to the proposed method was 66.6% that of the traversal method. Simulations and experiments fully demonstrate the feasibility of the method.

An Improved Back-Projection Algorithm for GNSS-R BSAR Imaging Based on CPU and GPU Platform

“…As can be seen, Target 25 is well focused. The range profile and the azimuth profile were well consistent with theoretical ones [23]. Table 3 shows the comparison of the proposed algorithm imaging and CPU imaging of target 25, the range and azimuth imaging resolution, Peak Side-Lobe Ratio (PSLR), and Integral Side-Lobe Ratio (ISLR) are the same.…”

“…The BPA can focus the echo data without the influence of geometry, while the frequency domain algorithms are greatly affected by the geometric structure [28]. Especially in the multisatellite fusion or multistation fusion mode [19,23], more complex geometry will lead to most of the frequency domain algorithm, which is no longer applicable. (2) The frequency domain algorithms rely on the accuracy of the equivalent squint model and cannot achieve a long-time synthetic aperture [41].…”

“…It can be seen that Figure 9a,b were very similar to each other-this shows that the image quality obtained by using GPU and CPU to run BPA is the same, which is consistent with the theoretical results. Based on resolution analysis [23], the theoretically predicted resolution is approximately 16.8 m in the range and around 0.95 m in the azimuth. The time taken was 128.04 s to run BPA imaging on the GPU, while that of the CPU was 19,974.24 s. The ratio of time consumed by GPU slightly higher than the simulation result; this is caused by I/O data transmission.…”

“…Secondly, the GNSS constellation has more than 140 satellites, which can provide more than 20 satellites in any certain area to observe the Earth's surface from multiple angles. Furthermore, abundant satellite resources provide the feasibility of fusing multiple satellites at different angles, not only to improve the spatial resolution and time resolution of the system but also to obtain more information on the characteristics of the observation area [23]. Finally, GNSS can provide precise timing services to ensure the synchronization performance of the technology.…”

Dual-Frequency Signal Enhancement Method of Moving Target Echoes for GNSS-S Radar

“…In references [14,15], the method of multiframe accumulation is used to accumulate GNSS echo signals for a long time, achieving the detection of maritime target echo signals. In references [16][17][18], a GNSS-based detection test for ground-level buildings was carried out. Since the target is still, long-term coherent integration is used to obtain the echo signal of the target.…”