摘要
VideoSAR成像时,运动目标发生散焦和移位,但在其真实位置留下阴影,且阴影在VideoSAR的图像序列中发生运动。根据运动目标的这个特征,提出了一种在VideoSAR图像序列中检测运动目标的方法,通过检测运动的阴影来检测运动目标。该方法主要分为背景补偿和背景分割两部分。首先,将段内的图像用SIFT+RANSAC算法进行配准,使得图像序列的背景对齐;然后,用单高斯模型对图像序列进行统计,得到该段图像序列的背景模型,将背景与当前图像差分后并二值化,得到前景的二值图;最后,对二值图进行形态学处理,即可提取出运动目标的阴影。该方法用美国Sandia国家实验室的VideoSAR图像进行了验证,具有良好的检测结果。
In VideoSAR mode imagery, the image of a moving target which is smeared and＇shifted, but the shadow of the target is displayed at the true target location and moves from frame to frame. Based on this characteristic of moving target, we present a novel technique for detecting moving targets in VideoSAR mode based on image processing techniques, using the clue of moving shadows to detect moving targets. This technique is mainly divided into two parts, background compensation and background segmentation. First, we register the image sequences with SIFT and RANSAC algorithms to compensate for the dynamic scenes. Next, we model the background through single Gaussian model（SGM）, and differentiate the background and current image to obtain the binary image of foreground. Then we extract the shadow of moving targets from the binary image by post morphological processing. This method has been demonstrated with real VideoSAR imagery sequence from Sandia Lab and has good detection result.
出处
《雷达科学与技术》
北大核心
2016年第6期563-567,573,共6页
Radar Science and Technology
基金
国家自然科学基金(No.61301212)
航空科学基金(No.20132007001)
试验技术研究项目(No.2015SY26A0003)
