摘要： 提出了一种基于灰度模板匹配与奇异值分解相结合的目标识别复合算法,并在目标搜索过程中采用了先粗匹配后细匹配的匹配方法,使得匹配速度和精确性大幅度提高;同时算法中消除了光线、目标大小变化等对模板匹配的影响;最后通过实验验证了该算法的实用性,取得了比较理想的实验效果。

关键词: 奇异值特征, 模板匹配, 目标识别

Abstract: A new multiplex method of Target Recognition based on the combination of template and singular value features is introduced. Themultiplex algorithm combines the two methods and complements each other, and the method of first rough then fine matching is adopted in theprocess of searching for the target, this makes the matching efficiency and accuracy be improved greatly. At the same time, the impact of the ray andsize of the target on the template matching is eliminated and the template can be updated real-time, which make the multiplex algorithm can bewidely applied. At last, the multiplex algorithm is validated to be very practical by experiment and get good experimental effect.

中图分类号: 

• TP391.4