研究了交通视频图像颜色和大小规一化方法,供后期动态目标的几何特征提取和模式匹配.由实验结果可以看出,经规一化处理的交通视频图像,效果比较理想.
In this thesis, the authors make a research on color and size normalization of the video surveillance for the geometrical characteristic and mode-matching of the dynamic objects in later period. The findings show that the effect of video surveillance by normalization is much better.