摘要
为解决焊接过程中由于高温高亮而使焊件变形难以测量的问题,采用改进的数字图像相关法和双目立体视觉对金属薄板焊接全场变形进行测量。对弱相关图像匹配的解决方法和数字图像相关的算法进行分析,并通过去噪和分段基准保证弱相关图像的正确匹配:首先,使用高斯平滑对原始图像进行去噪处理,消除焊接强光干扰;然后,采用分段建立基准的方法解决了焊点区域因散斑质量下降而出现的难以匹配的问题。对Newton-Raphson算法的二阶梯度矩阵进行简化,提高相关计算的速度。在测量实验中,首先对板料表面进行预处理与散斑制备,然后在氩弧堆焊(TIG)过程中利用高速相机拍摄板料焊接及冷却自由变形的全过程,最后计算所有图像中变形点的位移,获得板料特别是焊缝区的全场变形数据。研究结果表明:将此方法应用于焊接变形测量,能得到整个焊接过程中任意点的变形曲线,并且可在高温焊缝区获得准确的变形趋势。测量结果可用于分析影响焊接变形的因素,进而对焊接变形实施控制,也可用于检验焊接变形数值模拟方法的正确性。
It is hard to measure the deformation due to the high temperature and highlight during the welding process. Tosolve this problem, a method based on the modified digital image correlation and binocular stereovision technology wereused in the thin plate welding whole field deformation. The solution to match images of poor correlation and thealgorithm of digital image correlation were analyzed, and both the noise disposal and step benchmark were adopted toensure that the images of poor correlation could be matched correctly. Firstly, Gaussian smoothing was used to deal withthe noise of original image; then a method of setting up the piecewise benchmarks was adopted to overcome the matchingdifficulties which were caused by eliminate the effect of welding light, and the weak speckle performance around thesolder joint. In order to improve the speed of correlation calculation, the second order gradient matrix was simplified byNewton?Raphson algorithm. After the metal surface pretreatment and speckle preparation, some photos of tungsten insert gas (TIG) arc welding and cooling process were taken by the high speed cameras during the whole states of deforming inthe experiment. Eventually, the displacements of deformation points in all images could be calculated so that the data ofwhole field dynamic deformation on the thin plate especially the welding zone could be obtained. The results show thatthis method can get the strain curves of arbitrary points in the whole welding process, and the accurate deformation trendof the high temperature welding area can also be obtained. The measuring results can be used to analyze the influencefactors of welding deformation, control the welding deformation and verify the welding numerical simulation.
出处
《中南大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2017年第11期2936-2942,共7页
Journal of Central South University:Science and Technology
基金
国家自然科学基金资助项目(51675404
51421004)
广东省公益研究与能力建设专项资金资助项目(2014A010104003)~~
关键词
数字图像相关
双目立体视觉
薄板焊接
全场变形
分段基准
digital image correlation
binocular stereovision
thin plate welding
whole field deformation
piecewisebenchmark