基于特征图像方法的运动物体检测

在传统特征图像方法基础上,提出一种基于分块边缘特征图像的运动物体检测算法.该算法对背景图和叠加图进行分块处理,分别提取出背景图与叠加图的最大特征向量,求得背景图与叠加图的各块投影系数,进行运动物体的匹配检测.在实际处理时,采用了canny算子获取边缘图的方法来进行检测,以消除光照等外界干扰对检测的影响.

Motion Object Detection Method Based on the Characteristic Image

On the traditional foundation of characteristic image method, this paper proposes a kind of mobile examination algorithim based on the piecemeal edge characteristic image. This method carries on piecemeal processing to the background chart and the stacking chart, separately takes out their biggest characteristic vector, obtains each block projection coefficient, carries on the match examination of the mobile. In actual practice, the canny operator is used to obtain the edge detection so as to eliminate the external interference such as the detection of light.