基于支持向量机的拉弯工件弯曲质量检测研究

Research on bending quality detection of bending parts based on support vector machine

目前拉弯工件质量检测主要利用测量轮廓度,而轮廓度测量方法存在测量准确率与测量效率的矛盾。为此,利用机器视觉技术进行拉弯弯曲质量检测在保障准确率的同时极大地提高了效率。机器视觉技术处理过程如下,首先获取拉弯工件样本的投影视图,进行一系列的图像预处理,识别弯曲形状的内切圆,提取其圆心坐标和半径作为弯曲特征,利用支持向量机进行训练、预测。实验表明,其检测结果准确率达到100%且测量效率高。

At present,the quality measurement of bending parts is mainly used to measure the contour,and the contour measurement method has the contradiction between measurement accuracy and measurement efficiency.To this end,the use of machine vision technology for bending quality testing in ensuring the accuracy rate at the same time greatly improve the efficiency.The machine vision technology is as follows.First,the projection view of the drawing part is obtained,and a series of image preprocessing is carried out to identify the inscribed circle of the curved shape.The center coordinates and radius are extracted as the bending characteristics.The support vector machine is used to train and forecast The experiments show that the accuracy of the test results reached 100%and increased measurement efficiency.