大背景噪声下动目标实时检测方法研究

A Study on the Detection Method of Moving Targets in Large Background Noise

提出用帧间方块编码差值法和邻域综合法提取运动目标．第一种方法是把当前帧方块的二值图像像素编码值与前帧对应的二值图像像素编码值求差值，由差值检测运动目标；第二种方法是检测２×２方块内各像素的帧差值符号位和帧差绝对值，由二者判别是否运动像素．与普通帧差法相比，提高了抗噪度，改善了运动目标的边缘特性，减少了硬件量．检测算法可以以视频速率工作．实验结果表明，所提出的检测方法是有效的．

Real time detection of moving targets in large background noise is a rather difficult job. Two methods with moving targets to be extracted from the noise are proposed. In the first method, the difference between the pixels coding of the binary image of the block of present frame and that of previous frame is used to detect the moving targets. In the second method, the sign bit and the absolute value of the frame difference of every pixel in 2×2 block are detected and to see whether it is a moving pixel. In comparison to conventional method with frame difference, it is seen that the methods proposed are of better noise immunity, and with less hardware, while the edge performance of the moving object is improved. As compared with what proposed by T. Grafe and G. Scheffler (1988), the detection algorithms are simpler. The system can be operated at video rate with real time processing. The feasibility of two methods has been verified by experiments.