Improvement in computation time and image quality of fast ISAR image formation using SBR technique

Abstract

The target can be automatically recognized by analyzing the ISAR image of the target. It is called SAR ATR. SAR ATR requires a massive database of ISAR images of numerous CAD models for all observation angles. A fast ISAR image simulation method is necessary to rapidly build this multi-view database. And image quality is important issue because image quality affects the accuracy of target recognition. To quickly simulate ISAR image, Bhalla derived an image-domain ray-tube integration formula based on the SBR technique. This formula directly calculates the contribution of each ray to the ISAR image in the image-domain through a one-time ray tracing process at the central observation angle. Bhalla’s method, which is the previous fast ISAR imaging method, has a limitation that the ray tracing process must be performed at each central angle of the images. It is insufficient to rapidly build the multi-view database for SAR ATR because ray tracing process is performed at each central angle. To solve this limitation of previous method, fast ISAR image formation method over multi-aspect angles using the SBR technique was proposed. An enhanced image-domain ray-tube integration formula that calculates the ISAR image through the ray tracing process at an arbitrary angle, not at the central angle was derived in a closed form. Using this proposed formula, multiple ISAR image with different central angles can be generated through just one-time ray tracing process. Simulation results for a tank CAD model is used to verify the performance of the proposed method in this dissertation. Simulation results show that the proposed method reduces the computation time by 74% compared with Bhalla’s method. Bhalla’s method has another limitation that only FFT-based ISAR image is obtained. A FFT-based ISAR image has a restricted range resolution, and the scattering center has sidelobes in the FFT-based ISAR image domain. These restrictions make it difficult to analyze scattering properties of the CAD model. Thus, fast scattered field prediction method for super-resolution ISAR image formation was proposed in this dissertation. For this method, a FFT-based ISAR image formula is derived in a closed form. However, the field data obtained using the closed form formula have high error term. To suppress errors in the field data, the field truncation method is proposed. Simulation results for a helicopter CAD model and a tank CAD model is used to verify the performance of the proposed method in this dissertation. The field data and super-resolution ISAR image obtained by using the proposed method are in good agreement with the results for the conventional method, while the computation times are tremendously reduced. Especially, simulation results show that the field truncation method is essential for the proposed fast field prediction algorithm. In this dissertation, fast ISAR imaging methods to solve the two limitations of the previous fast ISAR imaging method using SBR technique were introduced. The first one is fast ISAR imaging method over multi-aspect angles using the SBR technique, and the second one is fast scattered field prediction method for super-resolution ISAR image formation using the SBR technique. Simulation results for complex PEC CAD models are used to verify the performance of the proposed methods.