中文摘要：

本文提出一种基于相位相关和差分相乘的动态背景下运动目标检测方法。首先用相位相关算法配准由于摄像机运动导致的图像背景旋转、缩放和平移量,然后用连续四帧图像差分相乘的方法分割出运动目标的轮廓,最后用形态学图像处理方法填补运动目标中的小空洞,并滤除残留的孤立噪声点。经过对多个视频序列的试验验证,结果表明该方法具有良好的性能,它对于各种因素的影响,如光照变化、背景中树叶晃动等局部微小变化、背景配准误差等,都有很强的适应能力。

英文摘要：

A method for moving object detection under dynamic background based on phase-correlation and differential multiplication is proposed in this paper. Firstly, rotation, scaling and translation of background caused by camera movement are matched using phase-correlation. Secondly, the outline of the moving object is segmented by multiplying the two difference images of four frames. Lastly, the small hollow in the moving object is filled, and the residual noise points are filtered by morphological image processing. The effectiveness and good performance of the method are verified by video sequence experiments. The proposed method has good adaptability to the influence of various factors, such as the illumination variation, partial small changes from leaf swaying in the background, background matching error, and etc.