结合分块的改进三帧差和背景差的运动目标检测

针对运动目标检测易受背景影响及帧间差分易产生空洞的问题, 提出了一种基于分块的改进三帧差分和背景差分相结合的运动目标检测算法. 该算法利用边缘检测法和均值法建立初始背景模型, 将视频图像划分成多个子块, 对利用改进的三帧差分和背景差分获取的图像的各个子块进行自适应阈值检测, 获取图像中的运动前景目标, 背景图像采取自适应更新方法. 实验结果表明, 该算法能完整的提取运动目标, 背景适应性强, 具有较高的准确性和效率.

Moving Target Detection Based on Improved Three Frame Difference and Background Difference

For the problems that moving target detection is affected by the background and inter-frame difference is easy to produce hollow. An improved algorithm is proposed for moving target detection. It is based on modular improved three frame difference with background difference. The initial background is established by edge detection and averaging method, then the video images are divided into multiple sub-block, detecting the video images sub-blocks which achieved in the use of improved three frame difference and background difference with adaptive threshold. To obtain the moving target in the image, the background image is updated by self-adaptive method. The experimental results show that the algorithm can extract the moving object completely, has high accuracy and efficiency, and the background has strong adaptability.