摘要
基于计算机泛视觉（ＣｏｍｐｕｔｅｒＰａｎ－Ｖｉｓｉｏｎ，简称ＣＰＶ）理论中借助非视觉信息辅助视觉问题求解的思想，研究了通过产品的ＣＡＤ模型提取视觉检测相关先验信息，从而自动生成图象匹配模板的方法。该方法利用ＤＸＦ文件中的实体Ｓｅｃｔｉｏｎ生成模板位图，并讨论了模板尺寸与匹配计算量的关系。提出一种基于模板图形特征的轮廓优先序贯匹配算法以减少匹配过程中的计算量。
According to the idea of computer pan vision theory that non vision information can be used to help to solve vision problem, method of generating match plate of inspected workpieces by means of prior information from CAD models of these workpieces is researched. This method forms the BMP image of match plate based on the entities section of workpiece's DXF file of AUTOCAD. The relation between the size of match plate and the complexity of match calculation is discussed. A configuration priority sequential match algorithm based on plate image feature is proposed to reduce the amount of match calculation. An example of the method in engineering application is introduced.
基金
高等学校博士点基金
