运动目标检测的三帧差法算法研究

Algorithm Research on Three-Frame Difference for Detection of Moving Target

运动目标因速度时变性强,路径不确定等特点给运动目标检测和跟踪带来难度。常用的运动目标检测方法虽然很多,但大都会因为噪声给检测的精度带来很大的影响。通过使用三帧差法提取目标,再用数学形态学的方法对提取出的图像进行处理,可有效检测出运动目标并消除噪声。实验结果表明,这种方法计算量小,受环境变化影响小,实时性强,目标检测准确性高.

Because of the features of moving target,such as strong speed degeneration and uncertain route,it is hard to detect and track the moving target.Although there are many detecting methods for moving target,the noise of detection affects the detecting precision dramatically.In this paper,we extract target by three frame subtraction,and then process the extracting image by the method of mathematical morphology.By this method,the moving target can be detected effectively and the noise is eliminated dramatically.The experimental results show that this method has less calculation,higher real-time performance and higher accuracy for moving target detection,and can be well applied indoors or outdoors.