A CFAR target detection method and system of SAR image based on multi-scale segmentation

Abstract

The invention discloses a CFAR target detection method and system of an SAR image based on multi-scale segmentation. Multi-scale segmentation is introduced into CFAR target detection, which makes fulluse of the advantages of multi-scale segmentation. A segmented image block is taken as a detection window, the homogeneity of the window internal background is ensured, the accuracy of clutter statistical modeling is enhanced, and the detection accuracy is improved. At the same time, the multi-scale segmentation can divide the heterogeneous background or the artificial object into a single imageblock, and eliminate most of the clutter in the image block through the determination of the adaptive threshold, thus suppressing the false alarm in the detection results. The method and the device can ensure the detection accuracy in practical application, restrain the generation of a false alarm target, and realize the fast target detection under the background of complex surface clutter of a large scene.

Description

technical field [0001] The present invention relates to the field of image target detection, more specifically, to a multi-scale segmentation-based SAR image CFAR target detection method and system. Background technique [0002] Synthetic Aperture Radar (SAR) is an active microwave remote sensing sensor. Compared with traditional visible light and thermal infrared remote sensing methods, it has all-weather and all-weather characteristics because it is not affected by weather such as clouds and rain. Compared with real aperture radar, synthetic aperture radar has higher azimuth resolution. Due to the excellent characteristics of SAR technology, more and more countries have developed their own SAR sensors, including Canada’s Radarsat2, Germany’s TerraSAR-X, Japan’s ALOS2 and China’s GF-3, accumulating a large amount of SAR image data. [0003] However, due to the coherent imaging characteristics of SAR itself, there are often a large number of coherent speckle noises on SAR i...

AI Technical Summary

Problems solved by technology

Traditional CFAR methods include CA-CFAR, GO-CFAR, SO-CFAR, OS-CFAR, and dual-parameter CFAR methods. This type of CFAR is based on sliding window detection, and the detection speed is slow and cannot adapt to a large-scale global situation. target detection

CFAR fast detection methods include global CFAR, two-level cascaded CFAR detection methods, and multi-layer CFAR detection methods, but these methods can only achieve marine target detection, and are not capable of detecting surface clutter in more complex scenarios.

Images