ISAR Real-Time Imaging System And Imaging Quality Evaluation

With the realization and development of 2d high-resolution inverse synthetic aperture radar imaging technology,the research on high-resolution ISAR real-time imaging technology is becoming more and more important.It is a very effective method to improve the range resolution of ISAR imaging with large bandwidth signal.However,because the range curvature of ISAR imaging with large bandwidth cannot be ignored,the traditional ISAR imaging algorithm will lead to geometric distortion.This paper uses frequency scaling algorithm to imaging,and the target velocity required for frequency scaling imaging is obtained by parameter estimation method.If only relying on the traditional method of serial computation in single-core or multi-core CPU for imaging,it will be difficult to deal with the huge amount of computation in a short time.Therefore,the parallel acceleration technology of radar imaging based on CPU+GPU heterogeneous platform came into being.It is also necessary to evaluate the image quality after ISAR imaging.To evaluate ISAR imaging,a set of evaluation method is proposed and implemented by software.The main works of the thesis are as follows:1.In this paper,the range bending problem in ISAR imaging with large bandwidth is studied.The parameter estimation method is used to solve the target velocity problem for frequency scaling algorithm.The imaging is improved by two minimum entropy self-focusing methods,one overall self-focusing and one Block self-focusing.Understand the parallelism of the algorithm from the derivation of the formula.2.Combined with the research on signal-level concurrency of each module in the imaging process,the parallel design of large bandwidth ISAR imaging algorithm is carried out by using the parallel development technology of CPU+GPU.CUDA technology is used to doing parallel programming simulating experiments.Focusing on the design of the parallel algorithm of the target velocity estimation,the idea and method of its accelerated realization are introduced in detail,so as to reveal the parallel design idea of the whole ISAR imaging.This method will make full use of GPU resources,configure threads,call more than 50 kernel functions,rationalize the memory usage in parallel execution to reduce the data transfering time and imaging time.Compared with the ISAR imaging algorithm that is executed serially on the CPU i7 dual processor with the main frequency of 3.4G alone,the parallel computation on the CPU+GPU heterogeneous platform of TITAN RTX graphics card with the same CPU can achieve an acceleration over 70 times.3.Starting from the quality assessment based on the point target model and the surface target model,a set of evaluation methods for ISAR image quality was developed and implemented in software.This method is suitable for the evaluation of imaging quality between simultaneous and the same type of IS AR imaging algorithms.