基于机器视觉的SOP贴片元件的定位检测方法研究

为了提高SOP类贴片元件的贴装效率,提出一种基于机器视觉的定位识别检测方法。针对SOP类封装贴片元件的成像特点,采用中值滤波代替Canny算子的高斯滤波进行图像边缘检测,以便达到更好的去噪效果;根据一定的轮廓周长和面积为限制条件剔除干扰轮廓和引脚轮廓的根部,只保留引脚足部的轮廓;通过计算中心矩的方法得到引脚足部的中心坐标,采用邻接矩阵和改进的最小二乘法相结合的方法计算出元件的偏转角度。该方法简单有效、运算量少、稳定性高。实验结果表明:该算法在保证SOP类元件高精度定位的同时,也能够实现高效率的贴装。

Research on Location Detection Method of SOP Patch Element Based on Machine Vision

In order to improve the mounting efficiency of the patch element of the SOP class, a method of location recognition and detection based on machine vision was proposed. Aimed at the imaging characteristics of the package mount element of SOP class,the median filter was used to detect image edges instead of the Gaussian filter of the Canny operator, so as to achieve better denoising effect.According to the limitation condition of certain contour circumference and area, the root of interfering contour and pin contour was removed, and only the contour of pin foot was retained. The center coordinate of the pin foot was obtained by calculating the center moment, and the deflection angle of the element was calculated by combining the adjacency matrix and the improved least square method. The method is simple and effective,with less computation and high stability.The experimental result shows that the algorithm can not only ensure high precision positioning of SOP class components, but also achieve high efficiency mounting.