High Range Resolution Profile Target Recognition Based On Sparse Representation

Radar Target Recognition plays an important role in the field of military. Highresolution range profile (HRRP) reflects the target geometric structure along thedistance and is convenient to be acquired. This dissertation studies the method of featureextraction and classification for the high resolution range profile target recognitionbased on Sparse Representation.The main content is summarized as follows:1. Based on radar scattering center model, as well as Sparse Representation theory,a HRRP target identification method using Gabor dictionary parameter for thecharacteristics is proposed, and simulations verify the effectiveness of Suchcharacteristics for HRRP target identification.2. When sparse representation algorithm is adopted to feature extraction for HRRP,a matched atomic library is constructed to reconstruct the radar data using the theory ofcompressed sensing. The consistency of the positions of the selected atoms in the libraryto the positions of the strong scatter centers is derived in thesis. The target strong scatterpoints location features in the process of reconstruction radar echo signal are extracted.The extraction of target length, and the relative position between scattering points areused as features to classification, And the characteristics of noise robustness is analyzed.3. A method for HRRP target identification on the basis of convex optimizationsparse representation theory is studied in this dissertation. Features such as PCA andLDA perform well for convex optimization sparse representation-based classification.Experiments verify the features of stability and robustness to noise to the sparserepresentation for recognition.