Discrimination Method of Radar Distributed Ground Targets Based on Nearest Neighbor Classifier

Abstract

The invention discloses a radar distributed ground target discrimination method based on neighbor one-class classifiers and mainly aims at solving the problems of superabundant false alarms after detection in the prior art. The radar distributed ground target discrimination method based on the neighbor one-class classifiers comprises the steps of 1) performing 2-norm normalization on radar one-dimensional echo signals and then performing constant false alarm detection, 2) intercepting and aligning the detection results and selecting training samples and a test sample, 3) calculating the average Hausdorff distance vector between the training samples and determining the judgment threshold Thr of the nearest neighbor one-class classifier, 4) calculating the average Hausdorff distance d between the test sample and the training samples, and 5) comparing the average Hausdorff distance d with the judgment threshold Thr, if d is less than or equal to Thr, identifying test sample as a target, otherwise, identifying test sample as non-target. The radar distributed ground target discrimination method based on the neighbor one-class classifiers is used for well removing non-target false alarms and can be applied to the discrimination of the distributed ground targets.

Description

technical field [0001] The invention belongs to the technical field of radar, relates to a target detection and identification method, and can be used for ground distributed target identification based on radar broadband echo. Background technique [0002] Since the 1960s, with the rapid development of radar technology, radar systems have been widely used in many fields such as military and civilian. According to the width of the instantaneous bandwidth of the radar transmission signal, radar can be roughly divided into narrowband radar and broadband radar. Commonly referred to as wideband radar means that the absolute bandwidth of the radar transmitted signal is relatively large. The main function of traditional narrowband radar is to locate and detect targets. With the wide application of radar technology, people expect to obtain more information about targets. Compared with traditional narrowband radar, wideband radar has the following advantages: [0003] One is that ...

AI Technical Summary

Problems solved by technology

Due to false alarms, there are non-target samples in the detection results. If the target recognition is performed directly after the detection, on the one hand, there may be non-target samples in the training samples, which will reduce the final target recognition performance. On the other hand, the calculation of the target recognition stage is very large. , the existence of too many non-target samples will cause a large amount of unnecessary waste of computing resources, and increase the cost required for the recognition stage

Images