Theory And Method Of Superresolution Imaging For Forward-looking Radar Of Moving Platform

In moving platform real beam scanning radar imaging, the achievable azimuth angular resolution cannot satisfy the increasing demands of practical applications, such as missile precise terminal guidance, terrain avoidance and self-landing. Although the current forward-looking mono-pulse beam sharpening technique can promote the visual effect of the imaging results, the resolution performance of the image cannot be improved. Therefore, it is of crucial practical value to study the moving platform forward-looking radar imaging method to improve the azimuth resolution and breakthrough the limitation of azimuth resolution of real beam radar.This dissertation studies the problems in moving platform forward-looking radar super-resolution imaging. The research work mainly include echo modeling, super-resolution imaging algorithm and airborne experiments.1. The quasi-convolution echo model is established for scanning radar, which provide theoretical support for the application of deconvolution methods in forward-looking super-resolution problems. Then, the signal to noise ratio constraint for super-resolution performance is studied to provide theoretical basis for the radar parameter determination in engineering applications.2. The super-resolution imaging algorithm for sparse targets is proposed, which can effectively restrain the noise and clutter amplification problems in deconvolution process and avoid false targets. The proposed method enhance the resolution performance of multiple isolated targets in noise and clutter environment.3. Based on generalized Gaussian distribution assumption for prior information of targets, the super-resolution algorithm for surface targets is proposed by tuning disperse-parameters to determine the objective functions. This algorithm can realize the super-resolution imaging for surface targets in noise and clutter environment and enhance the edge information for surface targets.4. The equivalent-array forward-looking super-resolution imaging algorithm is proposed to transform the forward-looking model into array model. Then, the array super-resolution imaging is realized based on least square and least mean square error criteria. The proposed method can accurately estimate the number and position of isolated targets under the condition of few scanning times.The above model and imaging methods are verified by using both simulation and real data. It is proved that the proposed super-resolution method can realize the super-resolution imaging for moving platform forward-looking radar.