基于支持向量机的磨削表面烧伤识别研究

Research on Grinding Surface Burn Degree Identification Based on Support Vector Machine

磨削加工是机械加工中的重要工艺之一,磨削烧伤对工件的各方面的性能有很大影响。因此,需要对磨削烧伤进行分级并识别以提高工件品质。论文采用支持向量机法对采集的磨削烧伤图像的颜色以及纹理特征进行深入地分析和研究。研究结果表明,随着数据量的变化,支持向量机模型对不同烧伤类型的识别准确率并不相同,当数据量为500时,支持向量机的类型识别准确率最高为94.02%。基本满足实际工程中对磨削烧伤识别的准确率要求,并具有一定的实际推广意义。

Grinding is one of the important processes in machining.Grinding burns has a great influence on the performance of all aspects of the workpiece.Therefore,it is necessary to grade and identify grinding burns to improve workpiece quality.In this paper,the support vector machine method is used to deeply analyze and study the color and texture features of the collected burned images.The results show that with the change of data volume,the support vector machine model has different recognition accuracy for different burn types.When the data volume is 500,the support vector machine has the highest accuracy of type recognition is 94.02%.It basically satisfies the accuracy requirement for the identification of grinding burn in actual engineering,and has certain practical promotion significance.