多光路下汽车齿轮磨削裂纹图像识别方法  

Image Recognition Method of Automobile Gear Grinding Cracks Under Multi-Optical Paths

作　　者：翟志方 胡超凡 马建平[2] ZHAI Zhi-fang;HU Chao-fan;MA Jian-ping(School of Automotive and Mechatronic Engineering Xinyang Vocational and Technical College,He’nan Xinyang 464000,China;School of Mechanical and Electrical Engineering Guilin University of Electronic Technology,Guangxi Guilin 541004,China)

机构地区：[1]信阳职业技术学院汽车与机电工程学院,河南信阳464000 [2]桂林电子科技大学机电工程学院,广西桂林541004

出　　处：《机械设计与制造》2023年第11期26-30,共5页Machinery Design & Manufacture

基　　金：2021年度广西壮族自治区科技厅自然科学基金项目—张量分析理论及其在高端机械装备故障诊断与预测方法研究(2021GXNSFBA075050)。

摘　　要：汽车齿轮磨削裂纹具有裂纹短、裂纹浅、不易发现等特点,在光学性质影响下存在成像光路多样、图像处理复杂等难点。因此,可以利用机器视觉方式对齿轮磨削裂纹进行图像识别,提出了多光路下汽车齿轮磨削裂纹图像识别方法。利用机器视觉的CCD工业相机采集汽车齿轮磨削裂纹图像,并对其进行滤波降噪和灰度化处理,获取无噪且色彩饱和度归零的图像数据。采用3D视觉传感器对预处理后的齿轮磨削裂纹图像数据进行分割,以分割后的像素点阈值区间为基础,提取齿轮磨削裂纹图像的形状、密度、面积特征。将齿轮磨削裂纹特征数据带入支持向量机,获取磨削裂纹图像的特征分类矩阵,根据特征分类矩阵实现齿轮磨削裂纹图像的识别。实验结果表明,所提方法识别容错性强、识别效果较好、识别时间较短,具有一定的实用性。Grinding cracks of automobile gears have the characteristics of short cracks,shallow cracks,and difficult to find.Under the influence of optical properties,there are difficulties such as diverse imaging optical paths and complex image processing.Therefore,the image recognition of gear grinding cracks can be carried out by machine vision,and an image recognition method of automobile gear grinding cracks under multi-optical paths is proposed.The CCD industrial camera of machine vision was used to collect the image of automobile gear grinding cracks,and the filter noise reduction and grayscale processed were performed to obtain the image data with no noise and zero color saturation.The preprocessed gear grinding cracks image data was segmented by 3D vision sensor,and the shape,density and area features of the gear grinding cracks image were extracted based on the pixel threshold interval after segmentation.The characteristic data of gear grinding cracks were brought into the supported vector machine,and the characteristic classification matrix of the grinding cracks image was obtained,and the gear grinding cracks image was recognized according to the characteristic classification matrix.The experimental results show that the proposed method has strong recognition error tolerance,high recognition efficiency and short recognition time,and has certain practicability.

关 键 词：机器视觉 汽车齿轮 磨削裂纹 滤波降噪 支持向量机 

分 类 号：TH16[机械工程—机械制造及自动化] TG61[金属学及工艺—金属切削加工及机床]