A New Method of Tank Detection for SAR Images

Abstract

The tank is one of important military targets. Tanks detection is the study focus of the synthetic aperture radar(SAR) image processing currently. But there may be many false alarms existed in the detection result with most of the traditional tank detection methods affected by the SAR speckle. A new method of tank detection for SAR images based on the features of SAR images is put forward by this paper. It uses Gauss low-pass filtering to smooth the original image and the geometric active contour model based on prediction theory metric to realize automatic segmentation. It detects candidate targets by removing little connected regions. Finally, it further removes the false alarm targets based on the grey characteristic of the shadow regions. Experimental results indicate that the method can effectively and rightly detect tanks of SAR images. Moreover, it is insensitive to the initial contour.