基于局部微分光流的运动对象分割

Moving Object Segmentation Based on Local Differential Optical Flow

运动对象分割是研究从场景的图像序列或视频中提取出运动目标的理论和方法,是计算机视觉中一个重要的研究方向,在军事和工业等领域有着广阔的应用前景。提出一种基于局部微分光流的运动对象分割算法。首先采用局部微分光流算法计算出场景的运动光流场并完成其初始分割,然后利用canny算子探测出对象的边缘信息并将其作为对光流场得到的运动信息的补充,从而分割出更为准确的运动对象。实验结果显示该方法具有良好的分割性能。

Segmentation of moving objects is a systematic method on the extracting moving objects from image sequences or video. It is an important research direction of computer vision and has a widely future in military and industrial fields. A moving object segmentation method based on local differential optical flow was presented. Firstly, a local differential optical flow algorithm was adopted to calculate the moving optical flow field and the initial segmentation was finished. Secondly, the canny operator was utilized to detect edge information of objects, which is used as the supplement of moving information obtained by optical flow field, so as to segment more exact moving objects. The experimental results demonstrate the good segmentation performance of the proposed approach.