高斯牛顿法图像配准的工件精确识别

李博; 陈燕; 邹湘军; 刘念; 蒋志林

doi:10.3788/OPE.20152313.0784

您当前的位置：

首页 >

文章列表页 >

高斯牛顿法图像配准的工件精确识别

更新时间：2020-08-13

- 高斯牛顿法图像配准的工件精确识别
- Accurate recognition of work-pieces based on Gauss-Newton image registration
- 光学精密工程 2015年23卷第10z期页码：783-789
- 作者机构：
  
  华南农业大学工程学院南方农业机械与装备关键技术教育部重点实验室,广东广州,510640
- 作者简介：
- 基金信息：
  
  广东省教育部产学研结合专项资金资助项目(No.2012B091000167)();广州市科技计划资助项目(No.201510010140)();广东省省级科技计划资助项
- DOI：10.3788/OPE.20152313.0784
  中图分类号： TP242.6;TP391.41
- 收稿日期：2015-06-02，
  
  修回日期：2015-06-20，
  
  纸质出版日期：2015-11-14
- 稿件说明：
移动端阅览
李博, 陈燕, 邹湘军等. 高斯牛顿法图像配准的工件精确识别[J]. 光学精密工程, 2015,23(10z): 783-789

LI Bo, CHEN Yan, ZOU Xiang-jun etc. Accurate recognition of work-pieces based on Gauss-Newton image registration[J]. Editorial Office of Optics and Precision Engineering, 2015,23(10z): 783-789
李博, 陈燕, 邹湘军等. 高斯牛顿法图像配准的工件精确识别[J]. 光学精密工程, 2015,23(10z): 783-789 DOI： 10.3788/OPE.20152313.0784.

LI Bo, CHEN Yan, ZOU Xiang-jun etc. Accurate recognition of work-pieces based on Gauss-Newton image registration[J]. Editorial Office of Optics and Precision Engineering, 2015,23(10z): 783-789 DOI： 10.3788/OPE.20152313.0784.

摘要

为提高机器视觉系统对工件的识别精度

提出在工件图像匹配时加入高斯牛顿图像配准算法

以使工件的识别精度达到亚像素级。首先

构建模板和目标之间的刚体变换模型

建立模板图像和目标图像的相似度残差项

利用Sobel算子对待检测图像的

方向和

方向分别进行卷积运算得到待检测图像的差分图;然后

利用该差分图计算待检测图像的雅可比矩阵

应用双线型插值法计算模板进行刚性变化后的变换矩阵;最后

运用高斯牛顿法寻找到使变换矩阵和目标区域相似度残差项达到极小值点的亚像素目标位置

从而使工件目标的识别精度达到亚像素级。实验结果表明

该方法对工件的识别精度为0.1 pixel

角度均差为0.05

满足了高精度识别工件的要求。

Abstract

To improve the recognition accuracy of a machine vision system for work-pieces

the Gauss-Newton algorithm was employed to the intelligent robot`s machine vision and the vision's recoginition accuracy was improved by sub pixel. Firstly

a rigid transformation model between template and target was built and the residual similarity between template and target image was established.Then Sobel algorithm was used for folded operation of the images detected in the

direction and the

direction to get the directional diagram of the images. The Jacobian matrix of the images to be detected was calculated using this directional diagram

as long as the transformation matrix was calculated using BI-linear interpolation after image rigid transformation. Finally

the target location of sub pixel where the residual similarity between template and directional diagram reach the minimum was detected using Gauss-Newton algorithms

and the recoanition accuracy was in sub-pixel. Experimental results show that the precision of work-piece recoanition is 0.1 pixel

which satisfies the requirement of high accuracy recognition on work-pieces.

关键词

Keywords

references

李振伟,张涛,孙明国. 星空背景下空间目标的快速识别与精密定位[J]. 光学精密工程,2015,23(2):590-599. LI ZH W,ZHANG T,SUN M G.Fast recognition and precise orientation of space objects in star background[J].Opt Precision Eng.,2015,23(2):590-599.(in Chinese)

王云飞. 基于SIFT和归一化互信息的SAR图像配准算法研究[D].西安:西安电子科技大学,2014,1. WANG Y F.Research on SAR Image Registration Arithmetic based on SIFT and NMI[D].Xi'an:Xidian University,2014,1.(in Chinese)

李平,魏仲慧,何昕,等. 采用多形状特征融合的多视点目标识别[J]. 光学精密工程,2014,22(12):3369-3375. LI P,WEI ZH H,HE X.et al..Object recognition based on shape feature fusion under multi-views[J].Opt. Precision Eng.,2014,22(12):3369-3375.(in Chinese)

林黄耀. MVS 在机械工件识别中的应用[J]. 工程技术,2013,427(6):191-194. LIN H Y.MVS`s application in the work-piece recognition[J].Engineering and Technology,2013,427(6):191-194.(in Chinese)

王彦, 傅卫平, 朱虹,等. 一种基于SIFT特征匹配的工件识别方法[J]. 西安理工大学学报,2009,25(2):203-206. WANG Y,FU W P,ZHU H, et al..A method for recognition of work-pieces based on SIFT characteristics matching[J].Journal of Xi'an University of Technology,2009,25(2):203-206.(in Chinese)

黄伟国,顾超,朱忠奎. 用于目标识别的PCA-SC形状匹配算法[J]. 光学精密工程,2013,21(8):2104-2110. HUANG W G,GU CH,ZHU ZH K.PCA-SC shape matching for object recognition[J].Opt. Precision Eng.,2013,21(8):2104-2110.(in Chinese)

林玉池,崔彦平,黄银国. 复杂背景下边缘提取与目标识别方法研究[J]. 光学精密工程,2006,14(3):510-514. LIN Y CH,CUI Y P,HUANG Y G.Study on edge detection and target recognition in complex background[J].Opt. Precision Eng.,2006,14(3):510-514.(in Chinese)

STEGER C.Similarity measures for occlusion,clutter,and illumination invariant object recognition[J].Pattern Recognition,2001,2191:148-154.

张彦峰,艾海滨,杜全叶,等.基于金字塔影像分割的水域提取在海岛礁空中三角测量中的应用[J]. 测绘通报, 2014(12):70-73. ZHANG Y F,AI H B,DU Q Y,et al..Application of water extraction algorithm based on pyramid image segmentation in automatic aerial triangulation for island and reef[J].Bulletin of Surveying and Mapping,2014(12):70-73.(in Chinese)

別术林. 基于互信息的医学图像配准算法研究[D].北京:北京交通大学,2014. BIE S L.Research of medical image registration method based on mutual information[D].Beijing:Beijing jiaotong university,2014.(in Chinese)

龚昌来,杨冬涛.一种改进的双线性插值图像放大算法[J]. 激光与红外,2009,39(8):899-901. GONG CH L,YANG D T.A novel bilinear interpolation method based on edge direction[J].Laser and In frared,2009,39(8):899-901.(in Chinese)

张小红,刘刚. 基于小波变换与双线性插值图像配比算法研究[J]. 江西理工大学学报,2011,32(1):2-4. ZHANG X H,L G.Research of image matching algorithm based on the wavelet transform and bilinear Interpolation[J].Journal of Jiangxi University of Science and Technology,2011,32(1):2-4.(in Chinese)

杨金宝. 基于灰度相似性测度的医学图像配准技术研究[D].济南:山东大学,2008. YANG J B.Medical image registration based on gray-level similarity techniques[D].Jinan:Shandong University,2008.(in Chinese)

于浩,陈雄,范晶晶. 高斯-牛顿法在基于能量的目标定位中的运用[J]. 计算机工程与应用,2007,43(27):124-126. YU H,CHEN X,FAN J J.The application of Gauss-Newton method based on target location of energy[J].Computer Engineering and Applications,2007,43(27):124-126.(in Chinese)

浏览量

421

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

暂无数据