基于单目视觉三维恢复算法的微观形貌原位测量方法研究  

In-situ Measurement Method of Microscopic Morphology Based on Monocular Vision 3D Recovery Algorithm

作　　者：安奕同 周平[1] 闫英 AN Yitong;ZHOU Ping;YAN Ying(Key Laboratory for Precision and Non-traditional Machining Technology of Ministry of Education,Dalian University of Technology,Dalian 116024,China)

机构地区：[1]大连理工大学精密与特种加工教育部重点实验室,辽宁大连116024

出　　处：《机械工程师》2022年第12期9-13,16,共6页Mechanical Engineer

摘　　要：工件表面微观形貌分析在机械加工及测量领域有着十分重要的作用。与二维形貌分析相比,三维形貌分析能够提供更全面的表面信息。然而,现有复杂、昂贵的先进测量手段始终无法实现原位测量。为了实现对工件表面现场检测和三维形貌表征,将基于单目视觉的SFS(shape from shading,从明暗恢复形状)算法及光度立体法应用至微观形貌原位测量领域,提出一种可以进行二维图像采集并进行三维形貌重构的原位测量方法,并对比两种算法在微观形貌三维恢复上的结果。基于上述方法,进行SiC工件表面三维形貌恢复试验,同时与共聚焦显微镜扫描结果比较。试验结果表明,相比于SFS方法,光度立体法重构结果更加准确,截线特征点高度差相对误差低于27.2%,截线二维参数相对误差低于32.8%,表面三维参数相对误差低于19.6%,为解决机械零件表面微观形貌原位测量难题给出了新的解决思路。Surface microform analysis of workpieces has a very important role in the field of machining and measurement.Compared with two-dimensional analysis,three-dimensional surface analysis can provide more comprehensive surface information.However,existing complex and expensive advanced measurement methods are not always available for in-situ measurement,and in-situ measurement is more attractive and valuable for research from the perspective of avoiding secondary processing and saving costs.To realize the in-situ inspection and 3D morphology characterization of workpiece surfaces,the monocular vision-based SFS(shape from shading)algorithm and photometric stereo method are applied to the field of in-situ measurement of microscopic morphology,and an in-situ measurement method is proposed that can perform 2D image acquisition and 3D morphology reconstruction,and the results of both algorithms are compared for 3D recovery of microscopic morphology.Based on the above method,a three-dimensional morphology recovery test is conducted on the surface of SiC workpieces,and the results are also compared with those of confocal microscopy scans.The experimental results show that the photometric stereo method has better reconstruction results than the SFS method,with the relative error of the height difference of the intercept line feature points lower than 27.2%,the relative error of the intercept line 2D parameters lower than 32.8%,and the relative error of the surface 3D parameters lower than 19.6%,which gives a new solution to the problem of in-situ measurement of the surface microscopic morphology of mechanical parts.

关 键 词：微观形貌 原位测量 SFS 光度立体法 

分 类 号：TH117[机械工程—机械设计及理论]