基于注意度的线激光视觉测量的三维形状检测

摘要

Considering the low recognition rate of the workpieces due to its placement orientation in production line,a curve shape recognition method of line laser vision measurement based on attention degree is proposed.Three-dimen-sion measurement of laser vision technology is used to extract the surface trace information,and then the orientation recognition is transformed into a curve shape recognition problem.In terms of the matching algorithms,two kinds of distance measurements are defined respectively as coarse matching and accurate matching of error evaluation function;As for recognition accuracy,a template similarity measurement based on attention degree is defined.Experimental re-sults show that the proposed method is of high recognition rate,short time-consuming,at the same time,insensitive to the workpiece translation and offset,which can fully meet the requirements of real-time detection in industrial produc-tion,and has wide applications.%针对工件在产线上的摆置方位识别困难识别率低等问题，提出了一种基于注意度的线激光视觉测量的曲线形状识别方法。采用激光视觉三维测量技术提取工件表面的光迹信息，将工件的方位识别转换为曲线形状识别的问题。在匹配算法方面，定义了两种距离测度，分别作为粗匹配和精确匹配的误差评价函数；在识别准确性方面，定义了一种基于注意度的模板相似性测度的分类方法。实验结果表明，本文提出的方法识别率高，耗时短，对工件的平移和偏移不敏感；能满足工业生产实时检测要求，应用范围广。