一种基于立体视觉的运动目标检测算法

在目标检测中采用立体视觉方法。首先对立体图像对进行匹配求取场景的视差图，再运用基于视差的背景差分法获得含有运动目标的前景区域，最后根据前景区域的视差和位置分布准确定位各运动目标。立体视觉方法有效解决了单目视觉检测方法中的一些难点问题，可以克服光线的变化和阴影干扰对目标检测带来的影响，在多个目标发生部分遮挡时仍能正确区分各运动目标。

A moving object detection algorithm based on stereo vision

A stereo vision method was used for detecting moving objects. First, the disparity map was got by using the stereo matching method. Then an algorithm based on disparity background difference was employed to detect the foreground region. Finally, every moving object was located correcdy according to the disparity and position of the foreground region. The stereo algorithm can solve the problems effectively which cannot be solved by the monocular method. By using the stereo algorithm, the effect on object detection caused by illumination changes and shadow can be eliminated, and the moving objects can be recognized correctly when they are occluded.