DRP CLASSIFIER FOR MULTI-CLASS TARGET RECOGNITION

In order to improve the recognition performance of the recognition algorithm for HRRP of multiple similar targets, we propose a multi-class target classification and recognition algorithm. By calculating the distance of each feature between the samples and calculating the ratio of the distances, the ratios of the distances of the different features are multiplied and finally the recognition result is obtained by voting. Since the main step consists of the calculation of distance, ratio and product, this classifier is called DRP classifier. In the paper, the DRP classifier is compared with other classification algorithms using five categories of target data with less feature differentiation, and the results show that the DRP classifier increases the accuracy of recognition and possesses higher accuracy than other classification algorithms when the target category increases.