基于帧差和小波包分析算法的运动目标检测

提出了一种在镜头不动的情况下基于累积帧差分割和小波包分析融合技术的运动目标检测方法.这种方法可分为四步:使用改进的累积帧差算法和阈值分割算法完成目标区域的分割,并获得初始运动模板;利用小波包分析算法提取出单帧图像的边缘信息并获得细化的目标区域边缘图;根据初始运动模板和空域边缘图像的融合得到更精确的运动目标模板;最后结合原序列图像检测出完整的运动目标.实验结果表明:这种方法可以有效地从对比度较小和噪声较大的视频序列中较精确地检测出完整的运动目标.

Moving Object Detection Method Based on Frame-Difference and Wavelet Packets Analysis

In this paper, propcse a unified moving object detection method based on accumulative frame- differences and wavelet packets analysis when an unmoving camera is used. The method includes the following four steps. Firstly, a rough motion mask is obtained by introducing a motion area segmentation method based on an improved accumulative frame- difference algorithm and optimal threshold segmentation. Secondly, the edges in each video frame are extracted using the wavelet packets analysis algorithm. Thirdly, an accurate motion object mask is obtained by integrating the rough motion mask and spatial edge image. Finally, the moving object can be extracted through combining the fine motion mask with the original image. The experimental results show that the proposed algorithm can extract moving object accurately from video sequences with low contrast and serious noise.