Methodology Of THz-ISAR Imaging And Its Realization On FPGA Chip

The terahertz(THz) region of the electromagnetic spectrum has proven to be one of the most potential and elusive part. Because of its short wavelength nature, the THz wave is very sensitive to surface and volume irregularities, and thus the combination of THz and inverse synthetic aperture radar(ISAR) is capable of obtaining much higher-resolution images, especially that of the non-cooperating targets. Moreover, with the encouraging potential capability in anti-stealth and anti-jamming, the THz-ISAR technology for space target imaging and recognition are of great theoretical and practical value.Field programmable gate array(FPGA) is a new programmable hardware device with high performance. The improvement of industrial manufacturing level has booted the integration of FGPA, and making the parallel processing for each module possible. FPGA has a superior performance in the application that has huge computation burden. Since the THz-ISAR requires a lot of computation, the efficient realization of the THz-ISAR methods on the FPGA is an appealing topic in the radar signal processing community.This thesis mainly investigates the critical image formation algorithms for THz-ISAR and its efficient realization on the FPGA chip. The author’s major contributions are outlined as follows:1. The principle and methodology for space target imaging using the THz-ISAR technology is investigated. The high speed motion of space target would de-coherent the radar echoes, broaden and translate the range profiles. Hence, an effective coherent method is investigated in order to improve the focus condition of the range profiles. By applying efficient envelop alignment and auto focusing methods, the translational motion error could be compensated, and the image geometry can be simplified into a turning-plate model. Since the space target always has large dimensions, range cell migration would occur for the scattering points. Therefore, 1-order Keystone transform is applied to realize migration correction, and then the high-resolution images can be obtained. The experimental results of simulated and real-measured data has proved the effectiveness of the proposed scheme.2. Based on the procedures of the THz-ISAR image formation process, the realization of auto-correlation method and Keystone transform on FPGA are explored. Comparative results between FPGA and Matlab simulation shows that the validity.3. During the image formation process, it is often required to deal with the echo data from the range and azimuth independently, which involves with the matrix transpose operation. Due to the limit resource of BRAM, the DDR2 SDRAM is often used to perform the matrix transpose. The large matrix is split into several sub-matrixes, and perform the transpose operation for each sub-matrixes independently using the RAM resource. At last, all the results are integrated. This can save the computation, RAM source, as well as the time greatly.