作者：项新建, 李超, 尤钦寅

作者单位：浙江科技学院自动化与电气工程学院, 浙江 杭州 310023

关键词：气门弹簧座;划痕检测;图像增强;傅里叶变换;双阈值

摘要：

为解决气门弹簧座内壁划痕检测人工检测精度以及效率不高的问题，实现自动化生产线，提出一种基于图像增强的气门弹簧座内壁划痕检测算法。针对弹簧座内壁图像背景复杂、噪声干扰大，采用最小二乘法拟合提取感兴趣区域，傅里叶变换滤波去噪抑制干扰。采用极坐标变换并基于划痕粗检测进行双阈值下的图像增强，实现划痕检测精准度和效率的提高。试验结果表明：通过传统直方图均衡化和全局图像增强对不同长度、方向的划痕进行检测，划痕检测准确率为76.83%，划痕长度准确率为87.86%；该算法能较好地抑制噪声，同时可提高气门弹簧座划痕检测的准确率，划痕检测准确率和划痕长度准确率分别为89.47%、93.62%，能满足实际检测需求。

Valve spring seat’s scratches detection algorithm based on image enhancement
XIANG Xinjian, LI Chao, YOU Qinyin
School of Automation and Electrical Engineering, Zhejiang University of Science and Technology, Hangzhou 310023, China
Abstract: In order to solve the problem of the low accuracy and efficiency of manual detection of scratches on the inner wall of the valve spring seat and realize an automated production line, an algorithm for detecting scratches on the inner wall of the valve spring seat based on image enhancement is proposed. Aiming at the complex background and large noise interference of the inner wall image of the spring seat, extract the region of interest by using least squares fitting, denoising and suppressing interference through Fourier transform filtering. Using polar coordinate transformation and the dual-threshold image enhancement based on rough scratch detection to improve the accuracy and efficiency of scratch detection. The test results show that:through traditional histogram equalization and global image enhancement to detect scratches of different lengths and directions, the accuracy of scratch detection is 76.83%, and the accuracy of scratch length is 87.86%; the algorithm can better suppress the noise and improve the accuracy of the valve spring seat scratch detection. The scratch detection accuracy and the scratch length accuracy are 89.47% and 93.62% respectively, which can meet the actual detection requirements.
Keywords: valve spring seat;scratch detection;image enhancement;Fourier transform;dual threshold
2023, 49(5):9-15  收稿日期: 2021-06-02;收到修改稿日期: 2021-08-11
基金项目:
作者简介: 项新建(1964-),男,浙江永康市人,教授,硕士,主要从事人工智能、机器人、物联网理论与技术研究
参考文献
[1] 张菁丽, 陈学文, 刘泽虎. 气门弹簧座多工位锻造工艺优化及数值模拟[J]. 锻压技术, 2012, 37(1): 34-37+59
[2] 张韵, 胡志远. 基于升降法测定气门弹簧座疲劳极限的试验研究[J]. 内燃机, 2019(1): 11-14
[3] 张涛, 刘玉婷, 杨亚宁, 等. 基于机器视觉的表面缺陷检测研究综述[J]. 科学技术与工程, 2020, 20(35): 14366-14376
[4] 胡晓彤, 董莹莹. 基于机器视觉的金属罐内壁缺陷检测[J]. 天津科技大学学报, 2014, 29(3): 63-67
[5] 谷明亮, 刘艳萍, 窦云朋, 等. 火电厂锅炉管道金属内壁缺陷检测方法研究[J]. 中国金属通报, 2020(9): 158-159
[6] 杨先凤, 赵玲, 杜晶晶. 改进中值滤波和形态学的油管裂纹检测算法[J]. 计算机仿真, 2018, 35(12): 81-85+180
[7] 曹义亲, 武丹, 黄晓生. 基于改进蚁群算法的轨道缺陷图像分类[J]. 计算机科学, 2019, 46(8): 292-297
[8] YUN J P, SHIN W C, KOO G, et al. Automated defect inspection system for metal surfaces based on deep learning and data augmentation[J]. Journal of Manufacturing Systems, 2020, 55: 317-324
[9] 罗兵, 任小洪, 李兆飞. 基于机器视觉的硬质合金微型喷嘴缺陷检测[J]. 机床与液压, 2021, 49(9): 115-120
[10] 田洪志, 王东兴, 林建钢, 等. 基于双阈值图像区域生长法的冲压件划痕检测[J]. 锻压技术, 2020, 45(6): 175-181
[11] 张德丰. 数字图像处理(MATLAB版)[M]. 北京: 人民邮电出版社, 2015: 381.
[12] 李克斌, 余厚云, 周申江. 基于形态学特征的机械零件表面划痕检测[J]. 光学学报, 2018, 38(8): 260-266
[13] 李长有, 陈国玺, 丁云晋. 改进Canny算子的边缘检测算法[J]. 小型微型计算机系统, 2020, 41(8): 1758-1762
[14] QI Y L, YANG Z, SUN W H, et al. A comprehensive overview of image enhancement techniques[J]. Archives of Computational Methods in Engineering, 2021, 29: 583-607