主动搜索耦合图案匹配的PCB 板基准点识别算法

摘要

The benchmark size change and the presence of a stain easily produce the problem such as high recognition, and in order to solve this problem,this paper put forward the active search coupling PCB reference point pattern matching recognition algorithm.First of all,the Gaussian median filter was introduced to smooth the image,the active search operator was designed by image segmentation threshold processing and morphology,the features of standard benchmark outline were searched,and the reference point of the first round of recognition was completed.Then,by benchmark template image and Aforge.NET matching function,and by locating datum targets in the image and then returning to the center coordinates and compatibility,the accurate identification of reference point was completed.Experimental results show that,compared with the current recognition technology,this algorithm has higher identification accuracy and anti-interference.%为了解决当前PCB板基准点识别算法在基准点尺寸变化和存在污痕的情况下易产生较高的误识别率等问题，提出主动搜索耦合图案匹配的PCB基准点识别算法。首先，引入高斯中值滤波对图像进行平滑处理，采用阈值处理与形态学分割图像，通过搜索符合特征标准的基准点轮廓，设计主动搜索算子，完成对基准点的首轮识别。然后，通过基准点模板图像和 Aforge.NET 匹配函数，在图像中定位基准点目标，返回中心点坐标与匹配度，完成对基准点的精准识别。实验结果显示：与当前识别技术相比，本文算法具有更高的识别精度与抗干扰性。