摘要
本文利用人眼对边缘(空间梯度)特别敏感的视觉特点,把帧间运动变化检测和图像的边缘检测结合起来,对帧间差分法进行了改进。首先,本文构造了一个二维运动探测器,将各向异性自适应边缘检测算子扩展到动态图像处理中,提出了,运动像素的能量差比较算法。该算法能够对运动目标各个运动像素的运动方向起到一个优化加强的作用,最终得到各个方向上的最佳检测。实验证明,这种算法可以很好的强化每个像素的运动方向,达到最佳的检测效果。
This article utilize the visual character that eyes are sensitive to the rim of an image （gratitude of space） ,we combine detection of movements between frames with detection of the rim of image to improve the frame difference algorithm. First, we create a two-dimension motion detector which extend the edge detection arithmetic operators of anisotropic adaptation into the processing of a dynamic image ,bring out a comparison algorithm based on energy deference of moving pixels. This method can strengthen moving object on every pixel according to every direction, finally, get prime detection with respect to every direction.
出处
《信号处理》
CSCD
北大核心
2007年第6期891-895,共5页
Journal of Signal Processing
关键词
能量差
各向异性边缘检测
运动目标检测
帧间差分
energy difference
anisotropic edge detection
motive target detection
frame difference
