基于块运动矢量分析的运动目标提取

Moving target extraction based on block motion vector analysis

为了从视频序列中提取完整的运动目标,提出运用块运动矢量和圆域选点法提取运动目标。首先,将相邻2帧图像按照相同的划分方式分成大小相同的块,通过计算每个块的运动矢量,去除只包含背景像素的块;然后,对剩余的块均进行再分块,通过块匹配准则进一步去除目标周边的背景块;最后,利用圆域选点法进行点配准,去除目标周边的背景像素点。通过上述步骤,可以得到完整的运动目标。研究结果表明:该方法对于处理复杂背景下的目标有很好的检测性,它不仅能够完整地提取出运动目标,而且提取出来的目标边缘清晰完整,目标中不出现空洞,检测效果好。

In order to extract the complete moving targets from video sequences, this paper proposed a method of using block motion vector and selecting point form the circle area to extract the moving targets. First of all, the two adjacent frames were divided into blocks with the same size by the same way, thus the pixel blocks which only contained background pixels were removed through calculating the motion vector of each block. Then, block the rest blocks again and further, removed the background blocks around the target through the block matching criterion. Finally, using the method of select point form the circle area to do point matching, the background pixels of the target were removed. Through the above steps, a complete moving target was got. The results show that this method has a good detection for processing target under the complex background. It can extract the whole moving target. The targets edges extracted are clear and complete, and the extraction of target does not include holes. So the testing effect is good. 8 figs, 20 refs.