运动检测算法的研究和仿真实现

Moving Target Detection Algorithm Research and Simulation

帧间差法和背景差法都是重要的运动检测方法,其核心问题在于如何得到准确的运动对象。针对该问题,本文提出一种结合帧间差和背景差的自动分割算法。该算法通过累积的帧差信息构建出可靠的背景,再将背景与当前帧比较,进而提取出视频运动对象。本文运用了最大类间方差法OTSU(又名"大津法")来获得自适应阈值,能更准确地对背景差图像进行阈值化分割,克服了传统固定阈值容易失效的问题。还采用了形态滤波的方法,对二值图像进行去噪,填充空洞。

Two consecutive frames subtraction and background subtraction are both important methods to detect the moving object. The core issue is how to obtain an exact moving object. So an automatic segmentation method algorithm is presented in this paper. Combing the advantages of temperal frames subtraction and background subtraction, effective background is modeled by accumulative frame-to-frame differences, and moving objects in video sequence could be extracted using background difference. OTSU segmentation method is presented, which segments the background difference image much more accurately, and overcomes conventional shortcomings of fixed threshold method. We also adopt mathematical morphological filtering, which can remove noise and fill in inanition.