Research And Development Of Broadband Polarization Radar Target Detection Algorithm

Modern battlefield environments get complicated day by day,which brings about huge challenges to precise guidance of radar.Traditional narrowband radar has been unable to meet operation requirements of complex ground scenes by degrees,and the wide-band system radar with a higher resolution ratio is needed for the image formation of the target to obtain richer target information.On one hand,the wide-band radar doesn't present the target as a target point any more,nor is the past CFAR algorithm applicable any more,the urgent demand for a wide-band radar with efficient target detection algorithm appears,and the development of a polarization system radar increases the polarization dimension information of a target,so it is an imminent task to do research on how to improve the performance of a detection algorithm by means of polarization information;on the other hand,The complexity of detection algorithm that matches the wide-band system radar rises to some extent,algorithm data increases,it is also of great significance to the precise guidance of radar.to implement the algorithm by the virtue of a high performance hardware platform in order to ensure the high real-time performance of the algorithm.For the demands above,this paper studies the polarized wide-band radar detection algorithm and its accomplishment respectively,and details are as follows:Research on classical CFAR algorithm for narrowband and wide-band radar targets.First of all,research on CFAR principles of several narrowband radar detection algorithms is conducted according to the model for the clutter of narrowband radars,different algorithms that apply to scenes are analyzed,and detection performances compared,thus the wide-band radar detection algorithm is derived,studies on the common energy accumulation detection algorithm and binary accumulation detection algorithm are done,reasons for the performance degradation of the two detection algorithms in different circumstances are analyzed,and detailed principles of the multichannel detection algorithm on the base of order statistics are analyzed in the background of Rayleigh distribution clutter,specific methods to get the detection threshold are given in the light of steps,and the accuracy of the algorithm and stability of its performance are verified by means of simulation and measured data in the end.Research on the detection algorithm for polarized targets of the wide-band radar is carried out.Firstly,several polarization expressions and their corresponding physical meanings of the radar are expounded.Polarization scattering matrix of a radar target is introduced,and polarization decomposition theory studied,among which focused studies on the statistical probability distribution of component products that resulted from Stokes polarization decomposition are implemented,being verified with measured clutter data.And on this basis,a polarized wide-band radar target detection scheme based on Stokes vector decomposition is put forward,and Stokes polarization decomposition applied to the original echo,among of which three polarized components are detected respectively with the multichannel detection algorithm on the base of order statistics,and detection results integrated to get final detection results.Finally,the algorithm performance is verified by means of measured data.The hardware experiment platform based on the TMS320C6678 DSP chip achieves the detection algorithm based on Stokes vector decomposition which is raised in the paper.Firstly,the hardware source of the chip is introduced,strengths and weaknesses of two main programming methods for the DSP chip are analyzed.Then,research on bare machine development process and specific steps is done,software architecture design for the detection algorithm introduced,algorithm optimization and compiler optimization are used to improve the real-time performance of the software on the basis of algorithm accomplishment.Finally,the accuracy of detection algorithm and enforceability of the project are verified through the contrast of MATLAB processing results to hardware processing results.