Research On Radar Target Recognition Based On High Resolution Range Profile

With the continuous change of high resolution radar and the characteristics of modernwarfare, radar can have the target type, true and false and attribute recognition function, to theeffective detection and warning. It can help target high-resolution reconnaissance and surveillanceon the battlefield, and to help the precision guided weapon precision hit the target, the field ofradar target recognition technique is emerge as the times require, and because of this, all over theworld, lots of scholars in the related research results continue to emerge. One dimensionalimage is more likely to be acquired and processed, can avoid the problems encountered in theprocess of image motion compensation.It is presently the most widely used method of radar targetrecognition.Based on the previous work about the radar target recognition, Firstly, this paper introducethe method for obtaining radar HRRP data, then studied a variety of target recognitionmethod, using some new methods, and made some improvement. All the method in this paper areproved effectively by experiment on real target data. Generally speaking, the main contents of thispaper include the following aspects:1. The first part study the characteristics of the radar HRRP. We discuss the radar scatteringcenter model and one-dimensional HRRP acquisition method, and points out that how to deal withthe target-aspect, time-shift and amplitude-scale sensitivity of HRRP samples is initially posedproblem for radar HRRP target recognition..Then a frame work for HRRP is established, whichlay the foundation for subsequent research.2. The second part is contributed to the target recognition method based on discriminantanalysis. We discuss some classical discriminant analysis method and kernel method such asLDA, KDA and then analysis LDA based on QR discrimination(LDA+QR) and KDA based onQR discrimination(KDA+QR). We points out the difficulty of calculate, and then improve themethod called Modified KDA based on QR discrimination (MDA+QR). The method keep highrecognition rate and at the same time, improve the computational efficiency.3. The third part focuses on the target recognition method and collaborative representationbased on sparse representation. We introduce sparse representation based on Fisher discriminationdictionary learning（FDDL）and target recognition based collaborative representation(CRC) into radar HRRP target recognition, and put forward a new method based on the latter called targetrecognition based kernel principal component and collaborative representation(KPCA+CRC). Therecognition rate, real-time in the method are improved, and show a clear advantage in use ofa large sample data.The two kinds of which researched in this paper are a classical subject and importantresearch area. Most of methods proposed are proved effectively by experiments on real target data,which improved in the recognition rate and real-time performance, have better pragmatic value.