基于图像拼接的电连接器插针检测方法研究

Research on Detection Method of Electrical Connector Pin Based on Image Mosaic

针对远心镜头无法调节视场大小,难以适应大尺寸矩形电连接器检测的问题,通过图像拼接的方法进行矩形电连接器测量与检测。将检测装置安装在工业机器人末端,通过机器人运动完成三帧待拼接图像的采集。对相机进行像素当量计算,采用中值滤波去噪,基于PatMax算法进行图像重叠区域的识别与匹配,引入Sobel边缘检测,完成图像的无缝拼接。通过设计测量算法,完成电连接器长度的测量、动态识别插针、定位插针、插针间距的计算。通过对2种型号的电连接器进行实验,结果表明:该方法对电连接器的尺寸测量平均相对误差低于0.1 mm,插针识别率达100%,拼接后插针间距与实际间距平均误差为0.02 mm,可满足实际应用中电连接器插针检测与测量要求。

In view of the problem that telecentric lens cannot adjust the field of view and is difficult to adapt to the detection of large rectangular electrical connectors,the rectangular electrical connectors were measured and detected by means of image splicing.The detection device was installed at the end of the industrial robot,and the three frames of the image to be stitched were collected through the robot movement.The pixel equivalent calculation of the camera was carried out.The median filter was used for denoising,the image overlapping area was identified and matched based on PatMax algorithm,and Sobel edge detection was in-troduced to complete the seamless stitching of the image.The measurement algorithm was designed to measure the length of elec-trical connector,identify pins dynamically,locate pins and calculate the spacing of pins.The experiments of two types of electrical connectors show that the average relative error of the method is less than 0.1 mm,the pin recognition rate is 100%,and the aver-age error of the distance between pins and the actual distance is 0.02 mm after stitching,which can meet the requirements of the detection and measurement of electrical connectors in practical applications.