动态背景下基于金字塔LK光流的运动目标检测

提出了一种动态背景下基于金字塔LK光流的运动目标检测方法。首先采用Shi-Tomasi角点检测法提取图像的强角点，再用金字塔LK光流法计算下一时刻这些特征点的对应位置，得到匹配的多组特征点对。利用这些特征点对建立仿射变换模型并计算仿射参数，对图像做仿射变换消除背景运动带来的误差，最后用帧间差分法处理得到前景目标。实验结果表明，本方法能够以较好的实时性、稳定地检测和提取动背景下的运动目标。

A Moving Object Detection Method Based on Pyramid LK Optical Flow under Dynamic Background

A moving object detection method based on pyramid LK optical flow under dynamic background is presented in this paper. Firstly, Shi-Tomasi corner detection method is used to extract the strong corners of the image, and then the corresponding position of these feature points is calculated by Pyramid LK optical flow method, and matched pairs of feature points are got. After that, these feature points are used to establish the affine transformation model and calculate the affine parameters to eliminate the error brought by the dynamic background. Finally, the frame-difference method is used to get the foreground target. The experimental results show that the method proposed in this paper can detect and extract the moving target in the dynamic background with better real-time stability.