Object trace point extraction method based on multi-dimensional fine characteristic analysis

Abstract

The invention discloses an object trace point extraction method based on multi-dimensional fine characteristic analysis. The method is mainly suitable for detection processing of an empty channel object by a three-coordinate radar. The implementation processes are as follows: trace point detection is carried out according to steps of distance condensation, grid division, grid communication and the like and stable track establishment is carried out; multiple fine characteristics of an object are calculated according to multicycle historical characteristics of a track object, comprising fluctuating characteristic, motion situation, object size, object clutter environmental description and object search region and the like; analysis comparison of the extracted object fine characteristics and detection trace points in the object search region is carried out, and whether object fine detection is needed is determined, and video data meeting the object amplitude modulation characteristic in the object search region is subjected to object parameter estimation. The algorithm is used for solving the problems of splitting of the same object point and wrong fusion of different object trace points during the trace point condensation process, such as situations that when a small object is close to a big object, the small object is fused with the big object, and when an object enters a clutter region, the object is fused with clutter.

Description

technical field [0001] The invention belongs to the field of radar target processing, in particular to a method for accurate detection and extraction of three-coordinate radar traces under the background of clutter. Background technique [0002] In the process of target detection and tracking, due to the problems of point trace splitting and point trace error fusion, the target is prone to filter trajectory deviation and tracking failure during tracking and filtering, so it is necessary to avoid point trace detection and extraction errors as much as possible. Through the statistical processing of various subtle features of multi-period targets, the dot trace information extracted by conventional detection methods is compared with the subtle features extracted from the target, and the search area where the target with poor correlation effect is located is re-detailed. Correct target dots. [0003] When the small target is close to the large target or the small target passes ...

AI Technical Summary

Problems solved by technology

[0003] When the small target is close to the large target or the small target passes through ground clutter, if the distance between the two is too close, the two will merge into one target. In this case, the small target will be deflected by the large target, resulting in small target tracking Failure; when the target track is split due to certain circumstances, the quality of the target track will be reduced or even the track will be split

[0004] Therefore, solving the problem of filtering failure caused by the detection and extraction error of the target is of great significance for improving the detection and tracking performance of the radar. The present invention is an algorithm designed for this problem

Images