摘要
针对数控机床加工领域刀具磨损区域的三维重构精度提高的问题,提出了变换域融合方法。将非下采样轮廓波变换和小波变换相结合,对非下采样轮廓波变换得到的低频系数采用小波变换方法进行融合,对得到的高频系数构造区域方差进行融合,突出刀具磨损图像的整体灰度信息和边缘细节信息,从而获得更为清晰的二维刀具图像,以及更为准确的磨痕形貌及磨痕体积。对基于非下采样轮廓波变换、小波变换方法和变换域融合方法得到磨损区域三维形貌和磨损体积与显微镜的观察结果进行了对比分析,发现提出的方法得到的三维形貌与显微镜结果最接近,且计算出的磨损体积平均相对误差小于10%。分析和实验数据表明,所提方法无论在图像的清晰度和磨损体积的准确性方面都有一定的优势,该方法对于刀具磨损区域三维重构精度较高,优于单独采用非下采样轮廓波变换和小波变换方法得到的结果。
To further improve 3D reconstruction accuracy of tool wear area,a transform-domain fusion method is proposed.Nonsubsampled contourlet transform(NSCT)and wavelet transform(WT)are combined in the transform domain.The low-frequency coefficients obtained by NSCT are fused with WT method,and the corresponding high-frequency coefficients are fused based on the regional variance criteria,so that the clearly-focused tool image is obtained to highlight the overall gray level information and edge detail information of tool wear image,so as to obtain clearer two-dimensional tool image and more accurate wear pattern.The 3D morphology and wear volume obtained by the three methods are compared with the results from microscope,it is found that this 3D morphology with the proposed method best coincides with the microscope result,and the average relative error of the wear volume is less than 10%.Compared with the traditional WT and NSCT methods,it is found that the proposed method has certain advantages in terms of the image clarity and accuracy of the wear volume and the reconstruction results are superior to ones obtained by using of non-subsampled contourlet transform or wavelet transform method alone for 3D reconstruction of the tool wear area.
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
2017年第12期76-83,共8页
Journal of Xi'an Jiaotong University
基金
陕西省自然科学基金资助项目(2016JM5054)
关键词
刀具磨损
变换域融合%三维重构
tool wear
transform-domain fusion
three-dimensional reconstruction
