论文部分内容阅读
通过双目视觉平台捕获图像反馈PC机,采用基于支持向量机(SVM)的机器学习方法对目标进行分类。判定型号并为每个型号加载图像处理检测方案及参数矩阵,实现电连接器针脚的柔性定位预处理,之后执行针顶轮廓的精确拟合算法,结合插针排列模版实现其三项检测工作。最后,通过实例验证与重复精度实验,结果表明本文方法具备面向多型号的柔性检测能力,并且稳定性,精度,效率满足低频连接器的检验要求。
PC was captured by binocular vision platform and classified by using machine learning method based on Support Vector Machine (SVM). Determine the model and load the image processing detection scheme and parameter matrix for each model to implement the flexible positioning pretreatment of the pins of the electrical connector and then carry out the precise fitting algorithm of the needle top profile and combine the pin arrangement template to achieve the three detection tasks. Finally, the experimental results show that the proposed method possesses flexible detection capability for multi-model, and the stability, accuracy and efficiency meet the inspection requirements of low-frequency connectors.