Ground static object classification method based on Hough transform

Abstract

The invention relates to a ground static object classification method based on Hough transform. The method comprises preprocessing of trace points of ground objects and using Hough transform for line extraction. A trace point distribution image of the ground objects is obtained via constant false alarm detection; a two-dimensional matrix is constructed for the trace point image according to distance-direction, and binarization is carried out; Hough transform is used to extract straight lines satisfying certain requirements from the image; and trace points associated with the detected lines in an image space are considered as interference objects and removed.

Description

technical field [0001] The invention belongs to the field of radar target detection and target recognition, and relates to a ground static target classification method based on Hough transform. Background technique [0002] In modern warfare, armed helicopters have become the most effective weapon to support ground forces in operations and seize ultra-low-altitude air supremacy and sea supremacy, and equipped with advanced airborne radar can greatly improve its combat power and combat effectiveness. An important task of helicopter-borne radar is to detect enemy targets such as ground tanks and armored vehicles. When the targets are in motion, they can be distinguished from ground stationary targets by MTD processing, but when they are in a static state, they will be overwhelmed In the echoes of many ground objects. [0003] When it is necessary to detect stationary armored targets, the echoes generated by ground buildings, high-voltage towers, roads, trenches, mountains, et...

AI Technical Summary

Problems solved by technology

[0003] When it is necessary to detect stationary armored targets, the echoes generated by ground buildings, high-voltage towers, roads, trenches, mountains, etc. will also be treated as targets. Although it is still interference clutter, it cannot pass through traditional MTI Wait for the Doppler filter to eliminate, which will cause the target to be eliminated together

At present, one-dimensional range images can be used for target classification and recognition. Theoretically reliable target classification methods can realize the extraction of armored targets, but in long-distance detection, the number of ground stationary targets that need to be identified is huge, and the amount of data storage and calculation are very large, it is difficult to meet the real-time requirements

Images