面结构光检测中符号识别方法研究  

Research of symbol recognition method of structured light detection

作　　者：喻英粽[1] 徐玉芳 喻擎苍[1] 

机构地区：[1]浙江理工大学信电学院,浙江杭州310018 [2]西安工程技术学院,陕西西安710065

出　　处：《计算机工程与设计》2010年第10期2339-2342,共4页Computer Engineering and Design

基　　金：国家自然科学基金项目(50575211)

摘　　要：为提高面结构光检测中符号识别效率和增强识别方法的鲁棒性,提出了一种基于形心特征面结构光符号识别方法。该方法采用逐行标记方法对图像进行标记,生成黑白连通域队列,接着对黑白连通域进行关联即可提取出待识别的符号域,根据符号域内连通数目、各连通成分形心分布特征、最大内连通域的方向和方位等特征即可实现符号的识别。整个识别过程只需扫描一遍图像和两遍连通域队列。To improve the efficiency of symbolic recognition and enhance the robustness of identification methods,a symbol recognition method based on the characteristics of the centroid is presented.This method uses progressive methods of marking tags on the images to produce black and white connected domain queue.Then the symbol domain can be extracted.According to the number of connected regions in the symbol domain,the connected components of the heart-shaped distribution,the position and orientation of the largest connected domain,symbol recognition can be achieved.The recognition process needed scanning the whole images once and connected domain queue twice.

关 键 词：符号识别 连通域 游程扫描 解码 形心 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]