汽车传感器电路板涂胶质量视觉检测系统研究

提出了基于视觉检测技术的车用传感器电路板涂胶质量视觉检测系统,采用多次模板匹配对目标图像进行定位和计算角度,并将图像平移和旋转,从而使零件图像与最初选定的图像位置和角度基本一致。然后,将从最初选定的图像获得的电路板应该有胶和无胶区域坐标应用到零件图像上,截取出相应区域的图像,利用图像分割等算法计算各个区域中涂胶部分的面积占总面积的百分比,与给定的面积百分比限制值进行比较,要求有胶区域的面积百分比大于对应的限制值,无胶区域的面积百分比小于对应的限制值则认为零件涂胶合格。结果表明,系统控制自动定位精度达到像素级,保证了检测精度,解决了目前在现场检测过程中人工操作工作量大、判断不准确以及效率低等问题。

Research on Vision Detection System Applied to the Circuit Board Gluing Quality of Automobile Sensor

Based on vision detection technology,the vision detection system of the circuit board gluing quality in automobile sensor is put forward. Multiple template matching is used in the position and angle calculation of the target image. The image are translated and rotated to realize that part image is consistent with the originally selected image on position and angle. Then,the glue area coordinate and without glueless area coordinate of circuit board obtained from the originally selected images is applied in parts image,the corresponding area images are subtracted to calculate the percentage of the total area on the glue area by using image segmentation algorithm,and compare it with the given percentage limit of the area. If the glue area percentage is greater than the corresponding limit and the glueless area percentage is less than the corresponding limit,then the glue part is considered to be qualified. The results show that the automatic positioning of system control achieves pixel level precision, can ensure the detection precision,can resolve the problems of the manual workload, inaccurate judgment and low efficiency in the detection process at present.