基于热像检测与分割技术的磨削砂轮堵塞检测研究  

Research on Grinding Wheel Loading Detection Based on Thermogram Detection and Segmentation Technology

作　　者：郭力[1] 江志顺[1] 尚振涛[1] 

机构地区：[1]湖南大学国家高效磨削工程技术研究中心,长沙410082

出　　处：《中国机械工程》2016年第5期577-583,共7页China Mechanical Engineering

基　　金：国家自然科学基金资助项目(51475157)

摘　　要：针对磨削加工中的砂轮堵塞检测的难题,提出了一种基于主动红外检测技术和热像图分割技术的砂轮堵塞检测方法。分别运用阈值分割、边缘检测和区域生长三种方法分割热像图,提取了砂轮堵塞量和状态的特征。实验结果表明:阈值分割法提取的砂轮堵塞磨屑的分布出现"偏聚"现象,堵塞特征受选取的阈值影响较大;边缘检测法克服了"偏聚"现象,但仍存在边缘不完整、定位不精确等问题;而基于梯度阈值过滤的区域生长分割法检测结果更客观和准确。把基于梯度阈值过滤的区域生长分割法的砂轮堵塞红外检测结果与显微照相法的砂轮堵塞检测结果进行了分析比较,结果表明基于梯度阈值过滤的区域生长分割的红外检测法可以实现对砂轮堵塞面积和状态的定量检测。According to the grinding wheel loading detection problems,a method was proposed for detecting the grinding wheel loading based on the active infrared detection technology and the thermogram segmentation technology.Three methods including the threshold segmentation,the edge detection and the region growing were used to segment the thermogram,the features like the amounts and states of the grinding wheel loading were extracted.The experimental results show that：the distribution of the grinding wheel loading swarf extracted by the threshold segmentation method presents a＂segregation＂phenomenon,and the selected threshold value is affected the loading features deeply.The edge detection method has overcome the＂segregation＂phenomenon,but there still exist some problems such as the edge is not complete and the positioning of the edge is not accurate.The results of the RGS method based on the gradient threshold filtering are more objective and accurate.The test result errors of the grinding wheel loading detection between the RGS method and the microscopic photography were given.The RGS method can realize active quantitative detection of the grinding wheel loading.

关 键 词：砂轮堵塞 红外检测 热像图 图像分割 区域生长分割法 

分 类 号：TG7[金属学及工艺—刀具与模具]