低对比度图像高精度轮廓三维重建算法研究  

Research on High Precision Contour Reconstruction Algorithms for Low Contrast Image

作　　者：杨武飞 蒋建军 陈雪梅 李中伟[2] 周钢[2] YANG Wu-fei;JIANG Jian-jun;CHEN Xue-mei;LI Zhong-wei;ZHOU Gang(Chengdu Aircraft Industrial(Group)Co.,Ltd,Chengdu 610000,China;School of Materials Science and Engineering,Huazhong University of Science and Technology,Wuhan 430074,China)

机构地区：[1]成都飞机工业(集团)有限责任公司,四川成都610000 [2]华中科技大学材料科学与工程学院,湖北武汉430074

出　　处：《软件导刊》2020年第6期223-226,共4页Software Guide

基　　金：中国博士后面上基金项目(2017M612453);国家重点研发计划高性能大型金属构件多电弧协同增材制造装备与工艺重点专项(2017YFGX)。

摘　　要：为了解决低对比度成像条件下传统三维轮廓重建算法精度偏低的问题,提出一种基于邻域距离最小化准则、由粗到细的高精度轮廓三维重建算法。首先使用Canny算子对目标图像进行边缘检测,再利用目标形状、尺寸等先验信息剔除分割出的虚假目标边缘。随后,计算边缘点法线方向上的梯度加权值作为工件初始化轮廓,根据左右一致性原则筛选出粗匹配点对,在匹配点对邻域附近使用亚像素进行精细搜索,利用邻域距离最小化准则估算出当前最佳亚像素匹配点对,并将其用于激光刻形工件的尺寸测量,总体精度可达0.1mm。该方法有效降低了轮廓重建算法对边缘定位精度的依赖,同时提高了轮廓重建精度。In order to overcome the drawbacks of low accuracy of traditional three-dimensional contour reconstruction algorithm in low contrast imaging,a high precision version based on the neighborhood distance minimization criterion and coarse-to-fine strategy is pro⁃posed.Firstly,the Canny operator is used to detect the edge of the target image,and then the priori information like the shape and size of the target is used to remove the fraudulent target edge.Then,the gradient weighted value in the normal direction of the edge points is calculated as the initial contour of the work piece.The rough matching points are selected by the left-right consistency principle.The sub-pixels are searched in the matching points,and the optimal sub-pixel matching points are estimated by the criterion of minimizing the neighborhood distance.It can be used to measure the dimension of laser engraving work piece with an overall accuracy of 0.1 mm.The algorithm effectively reduces the dependence of contour reconstruction algorithm on the accuracy of edge location and improves the accuracy of contour reconstruction.

关 键 词：低对比度图像 轮廓三维重建 邻域距离最小化准则 尺寸测量 激光刻形 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]