轮廓匹配方法改进及其在转子中的应用
PDF下载 (239)王 壮,戚嘉辉,赵一鸣,胡燕海.轮廓匹配方法改进及其在转子中的应用[J].宁波大学学报(理工版),2022,35(4):34-39.DOI:
WANG Zhuang,QI Jiahui,ZHAO Yiming,HU Yanhai.Improvement of contour matching method and its application in rotors[J].Journal of Ningbo University(Natural Science & Engineering Edition),2022,35(4):34-39.DOI:
| Title: | Improvement of contour matching method and its application in rotors |
| 作者: | 王 壮, 戚嘉辉, 赵一鸣, 胡燕海 |
| Author(s): | WANG Zhuang, QI Jiahui, ZHAO Yiming, HU Yanhai |
| 关键词: | 轮廓提取; 轮廓匹配; 特征描述子; 特征点匹配; 仿射变换 |
| Keywords: | contour extraction; contour matching; feature descriptor; feature point matching; affine transformation |
| 分类号: | TH-39 |
| 文献标识码: | A |
| 摘要: | 汽车发动机转子装配误差容易造成零件磨损和效率降低. 针对转子零件装配精度要求, 提出一种基于轮廓与特征点结合的轮廓匹配方法. 首先对转子图像进行滤波、二值化预处理; 其次提取轮廓并利用轮廓Hu矩方法进行轮廓粗匹配; 然后提取特征点, 并在以特征点为中心、半径为16像素的圆形区域内生成4个特征描述子作为匹配特征, 利用欧氏距离进行特征点匹配; 最后通过找到的匹配对计算仿射变换矩阵, 得到精匹配结果. 实验结果表明, 该方法在背景复杂、光照变化的条件下能够较精确地匹配轮廓, 保证匹配精度在1像素内, 且生成的特征描述子对数越多, 匹配结果越精确. |
| Abstract: | Vehicle engine rotor assembly error is easy to cause parts friction and efficiency reduction. Aiming at the assembly precision requirements of rotor parts, a contour matching method based on the combination of contour and feature points was proposed. Firstly, the image is preprocessed by filtering and binarization. Secondly, contour matching was carried out by contour Hu moment method. Then the feature points were extracted, and four feature descriptors were generated in the circular region with the radius of 16 pixels and the feature points were matched by Euclidean distance. Finally, the affine transformation matrix is generated by finding the matching pair, and the precise matching result is obtained. The experimental results show that the proposed method can match the contour accurately under the condition of complex background and light change, and ensure the visual detection accuracy within one pixel. The more the number of feature descriptors generated, the more accurate the matching result. |
| 参考文献 /References: | [1] 魏中雨, 黄海松, 姚立国. 基于机器视觉和深度神经网络的零件装配检测[J]. 组合机床与自动化加工技术, 2020(3):74-77; 82. [2] 童小彬, 孟婥, 孙以泽, 等. 基于HOG特征描述的轮廓匹配算法[J]. 东华大学学报(自然科学版), 2020, 46(5):787-792. [3] 尚小强, 陈从桂, 欧阳光. 基于机器视觉的轮廓度检测算法研究[J]. 机电工程技术, 2019, 48(1):21-23. [4] Korman S, Reichman D, Tsur G, et al. Fast-match: Fast affine template matching[J]. International Journal of Computer Vision, 2017, 121(1):111-125. [5] 杨白雪, 张晓丹. 尺度不变特征变换的图案匹配研究[J]. 国外电子测量技术, 2020, 39(2):53-58. [6] Chen X N, Zhao J, Yang R F. Product appearance detection based on visual keywords matching[J]. Microprocessors and Microsystems, 2020, 76:103062. [7] 王琼, 袁建英, 李柏林. 一种融合特征点与轮廓信息的匹配算法[J]. 计算机应用研究, 2014, 31(10):3145-3147. [8] 刘屿, 孙坤, 谢宏威, 等. 基于形状上下文特征和ICP的高精度轮廓视觉检测算法[J]. 华南理工大学学报(自然科学版), 2018, 46(1):131-138; 144. [9] Pallotta L, Giunta G, Clemente C. Subpixel SAR image registration through parabolic interpolation of the 2-D cross correlation[J]. IEEE Transactions on Geoscience and Remote Sensing, 2020, 58(6):4132-4144. [10] 党士许, 李磊, 张志鸿. 基于改进亚像素边缘提取的一种异性纤维检测方法[J]. 计算机与数字工程, 2016, 44(4):596-600; 682. [11] Sun X, Shang K, Ming D L, et al. A biologically-inspired framework for contour detection using superpixel-based candidates and hierarchical visual cues[J]. Sensors, 2015, 15(10):26654-26674. [12] 江波, 徐小力, 吴国新, 等. 轮廓Hu不变矩的工件图像匹配与识别[J]. 组合机床与自动化加工技术, 2020(9):104-107; 111. [13] Sun H F, Yang J H, Fan R B, et al. Stepwise local stitching ultrasound image algorithms based on adaptive iterative threshold Harris corner features[J]. Medicine, 2020, 99(37):e22189. [14] 朱思聪, 周德龙. 角点检测技术综述[J]. 计算机系统应用, 2020, 29(1):22-28. [15] 李东, 万贤福, 汪军, 等. 基于轮廓曲率特征点的服装款式识别方法[J]. 东华大学学报(自然科学版), 2018, 44(1):87-92. [16] Khurana M, Singh H. Data computation and secure encryption based on gyrator transform using singular value decomposition and randomization[J]. Procedia Computer Science, 2018, 132:1636-1645. |
| 备注/Memo: | 收稿日期: 2021－06－03. 宁波大学学报（理工版）网址: http://journallg.nbu.edu.cn/基金项目: 国家自然科学基金（51705263）.第一作者: 王壮（1997－）, 男, 山东潍坊人, 在读硕士研究生, 主要研究方向: 机电液产品开发. E-mail: 958563063@qq.com*通信作者: 胡燕海（1966－）, 男, 浙江宁波人, 教授, 主要研究方向: 机电液产品开发、生产计划与控制等. E-mail: huyanhai@nbu.edu.cn 宁波大学学报（理工版）网址：http://journallg.nbu.edu.cn/ |
