摘要
针对实时性要求较高的跟踪印花系统,提出了一种利用图像灰度特征进行跟踪匹配的算法.新方法首先通过比较直方图特征进行粗匹配,挑选出模板存在可能性较大的区域,然后通过计算横向与纵向的投影特征来实现模板的精确匹配.实验结果表明,相对于传统的灰度互相关算法,新算法匹配效果更为精确,同时减少了计算量,能够很好地满足印花系统高实时性的要求.
Demanding for real-time tracking of printing system,a template matching algorithm based on the image gray characteristic was presented.Firstly the new method did a rough match by comparing the histogram feature,so as to pick up the regions that the target may probably exit,secondly it did an accurate match by calculating the horizontal and vertical projection.The experimental results show that the new template matching method is much more accurate compared with the traditional gray cross-correlation algorithm,and it also reduces the amount of operation obviously,so the new algorithm could satisfy the high real-time demands of the printing systems well.
出处
《机械与电子》
2010年第9期78-80,共3页
Machinery & Electronics
关键词
模板匹配
直方图特征
投影
高实时性
template matching
histogram feature
projection
high real-time
