建筑机械加工轴承表面缺陷光学识别模型  

Optical Recognition Model of Bearing Surface Defects in Building Machining

作　　者：焦宏涛[1] 赵嵩[2] JIAO Hongtao;ZHAO Song(Zhengzhou Railway Vocational&Technical Colleg,He’nan Zhengzhou 450052,China;He’nan University,He’nan Zhengzhou 450046,China)

机构地区：[1]郑州铁路职业技术学院,河南郑州450052 [2]河南大学,河南郑州450046

出　　处：《机械设计与制造》2025年第1期214-217,221,共5页Machinery Design & Manufacture

基　　金：河南省科技厅科技攻关项目(212102210281)—基于AI的SLAM智能机器人关键技术的研发。

摘　　要：建筑机械加工轴承表面缺陷识别技术的表面缺陷识别效果不佳,影响工业生产的安全性。为了解决这一问题,提出建筑机械加工轴承表面缺陷光学识别模型构建方法。获取多角度机械加工轴承表面图像,将建筑机械加工轴承表面展开,再进行二维图像拼接,获得建筑机械加工轴承表面没有重复且完整的二维图像;通过局部与全部相结合的平滑策略构建光流误差轨迹模型,在光流求解策略的基础上,构建建筑机械加工轴承表面缺陷光学识别模型,计算建筑机械加工轴承表面二维图像的光流值,根据计算结果对建筑机械加工轴承表面缺陷情况进行判定,实现建筑机械加工轴承表面缺陷的识别。实验结果表明,所提方法的轴承表面缺陷识别率高、识别用时明显提升。The surface defect identification technology of bearing machined by construction machinery has poor identification ef⁃fect,which affects the safety of industrial production.In order to solve this problem,the construction method of optical recogni⁃tion model of bearing surface defects in construction machinery is proposed.Obtain the multi angle machining bearing surface im⁃age,unfold the building machining bearing surface,and then mosaic the two-dimensional image to obtain the non repeated and complete two-dimensional image of the building machining bearing surface.The optical flow error trajectory model is constructed through the smoothing strategy of the combination of local and all.On the basis of the optical flow solution strategy,the optical flow identification model of bearing surface defects of building machining is constructed,the optical flow value of twodimensional image of bearing surface of building machining is calculated,and the defects of bearing surface of building machin⁃ing are determined according to the calculation results.Realize the identification of bearing surface defects in construction ma⁃chining.The experimental results show that the recognition rate of bearing surface defects is high and the recognition time is sig⁃nificantly improved.

关 键 词：机械加工 轴承表面 缺陷识别 图像展开 图像拼接 

分 类 号：TH16[机械工程—机械制造及自动化] TH133.331