A Range Perturbation Approach for Correcting Spatially Variant Range Envelope in Diving Highly Squinted SAR With Nonlinear Trajectory

KT and azimuth sub‐region deramp‐based high‐squint SAR imaging algorithm mounted on manoeuvring platforms

“…Therefore, the number of FLOPs of azimuth compression can be calculated by ANCS is widely used in the frequency‐domain SAR imaging algorithms for azimuth compression [11, 13, 19, 20]. In general, ANCS [19], which is also used in [20], requires three azimuth FFT operations and three complex multiplication operations. Using the parameters in Table 1, the computational complexity of the proposed azimuth compression method is about 60% of that of the ANCS method.…”

“…Therefore, can be approximated to 1.8 azimuth FFT operations and 1.8 complex multiplication operations, given by In this case, the number of FLOPs of the proposed algorithm is Similarly, the number of FLOPs of the reference algorithm [20] is Compared to (43) and (44), we can find the computation of the proposed algorithm is slightly higher than that of the frequency‐domain imaging algorithm based on ANCS processing due to a large computing requirement for KT. For the echo signal generated by the parameters in Table 1 with 6000 sampling points in azimuth and 5120 sampling points in range, the computational times of the reference algorithm [20] and the proposed algorithm are about 8.0 and 11.8 s, respectively, using the personal computer with 2.7 GHz central processor unit clock frequency and 8 GB random‐access memory. The ratio of two computational times 11.8/8.0 (1.47) is slightly higher than the theoretical value (1.40) owing to some extra computations on the azimuth sub‐region deramp.…”

“…To verify the performance of the proposed algorithm, simulation experiments are carried out in this section. For high‐squint SAR imaging mounted on manoeuvring platforms, introducing the perturbation function is an effective way to correct spatial‐variant RCM [19, 20]. To present the advantage of KT on spatial‐variant RCMC, Dang et al [20] based on perturbation function and ANCS is selected as the reference algorithm.…”

“…To remove azimuth‐dependent RCM, Li et al [19] correct the first‐order azimuth spatial variabilities of both second‐ and third‐order RCM terms by introducing higher‐order perturbation factors in the azimuth time domain. Considering the range‐dependent squint angle, Dang et al [20] correct first‐ and second‐order azimuth spatial variabilities of range curvature by introducing third‐ and fourth‐order filter factors in the 2D frequency domain. The above algorithms make great achievements in correcting azimuth‐dependent RCM, but they cannot correct range–azimuth coupling RCM, which limits the extension of the imaging region.…”

“…In azimuth compression processing, one by one range of cell processing can effectively solve the range‐dependent azimuth focusing parameters. For the azimuth dependence, many algorithms adopt the method of azimuth non‐linear chirp scaling (ANCS) [16, 19, 20], which can only correct the first‐ and second‐order spatial variabilities of Doppler rate and first‐order spatial variability of third‐order Doppler parameter. When the azimuth width of the scene is large, the residual high‐order spatial variabilities of Doppler rate and high‐order Doppler parameters will significantly affect the focusing quality of the scattering points located in the azimuth edge of the scene.…”