数字视频分析中快速边缘检测和运动估计研究

本文研究了将粒子群优化算法同万有引力算法相结合进行边缘检测,利用万有引力原理进行启发函数的计算,指导蚁群运动趋向,快速检测出边缘线,作为图像分析的预处理结果。另外根据视频图像中局部运动集中在图像中部概率较大的特性,提出对图像的几个小区域数据进行全局运动估计,这样大幅度减少了运算量,实验中通过采样四个六分之一原图边长的矩形区域进行全局估计,实验证实了在精度相同的情况下,运算速度提高了九倍左右。

Research Fast Edge Detection and Motion Estimation in Digital Video Analysis

This paper first study the edge detection based on particle swarm optimization algorithm and universal gravitation algorithm in this paper. Using the principle of universal gravitation to calculate the heuristic function, guide the ant colony movement trend, and quickly detect the edge of the line. The detection results can be used as the preprocessing results of image analysis. We propose a global motion estimation based on several small domain data of the image according to the video of the local image motion is concentrated in the central part of the image of greater probability characteristics. In this way the amount of computation can be greatly reduced. We sampled four of 1/6 rectangular area original length in the experiment of global estimation. The experiment proves that the operation speed is increased by about nine times in the same accuracy.