第14卷第2期 智能系统学报 Vol.14 No.2 2019年3月 CAAI Transactions on Intelligent Systems Mar.2019 D0:10.11992/tis.201712024 网络出版地址:http:/kns.cnki.net/kcms/detail/23.1538.TP.20180507.0952.002.html 金属腐蚀区域图像增强算法研究 雷芳,熊建斌2,张磊,郭斯羽 (1.广东石油化工学院计算机与电子信息学院,广东茂名525000,2.广东技术师范学院自动化学院,广东广 州510665:3.湖南大学电气与信息工程学院,湖南长沙410082) 摘要:针对金属腐蚀区域图像中存在暗细节对比度不高、光照不均匀及颜色特征需保护的问题,提出一种在 HSI模型下的多尺度细节自适应增强与同态滤波的增强算法。首先,对RGB腐蚀图像进行色彩空间变换,保 留其中的色调和饱和度分量不变,对亮度分量进行增强。然后,通过小波变换进行多尺度细节自适应增强,提 升细节对比度并作分块同态滤波,改善光照不均的影响,获得增强后的腐蚀图像。实验结果表明,所提方法增 加了腐蚀暗细节的对比度,提高了金属腐蚀区域图像的整体亮度并保证了色彩信息的不失真。 关键词:金属腐蚀图像:HSI模型:多尺度:细节自适应增强;分块同态滤波 中图分类号:TP391文献标志码:A文章编号:1673-4785(2019)02-0385-08 中文引用格式:雷芳,熊建斌,张磊,等.金属腐蚀区域图像增强算法研究.智能系统学报,2019,14(2):385-392. 英文引用格式:LEI Fang,XIONG Jianbin,.ZHANG Lei,.etal.Image enhancement algorithm in metal corrosion areaJ.CAAl transactions on intelligent systems,2019,14(2):385-392. Image enhancement algorithm in metal corrosion area LEI Fang',XIONG Jianbin',ZHANG Lei',GUO Siyu' (1.College of Computer and Electronic Information,Guangdong University of Petrochemical Technology,Maoming 525000,China; 2.School of Automation,Guangdong Polytechnic Normal University,Guangzhou 510665,China;3.School of Electrical and Inform- ation Engineering,Hu'nan University,Changsha 410082,China) Abstract:Considering the images of the metal corrosion areas,the dark details have low contrast,and the illumination is not uniform;meanwhile,the color characteristics need to be preserved.To solve these problems,an approach based on multi-scale detail-adaptive enhancement and homomorphic filtering is proposed on the basis of the HSI model.First,the RGB corrosion image was color-transformed,whereby the hue and saturation components were preserved,and the lu- minance component was enhanced.Then,wavelet transform was used to implement multi-scale detail-adaptive enhance- ment,increase the contrast of the detail,and apply block homomorphic filtering,so as to improve the impact of non-uni- form illumination.Consequently,the corrosion image was enhanced.Experimental results show that the proposed meth- od can increase the contrast of dark details and improve the overall brightness of the image of the metal corrosion area, ensuring that the color information is undistorted. Keywords:metal corrosion image;HSI model;multi-scale;detail-adaptive enhancement;block homomorphic filtering 金属材料受到环境因素的影响其腐蚀表面呈型-。在腐蚀无损检测的远程视觉检测(remote 现形貌纹理、灰度、颜色等不同特征。通过对这 vision inspection,RV)检测中,主要是使用视觉设 些特征的分析可识别材料腐蚀的程度和腐蚀的类 备来获取腐蚀信息。对于金属外部腐蚀,可以采 用摄像机或CCD相机获取腐蚀图像,设备内部腐 收稿日期:2017-12-26.网络出版日期:2018-05-09. 基金项目:国家自然科学基金项目(61473331,61471167,61571147): 蚀图像的获取主要有管窥镜、自动爬行器等设 茂名市科技计划工业攻关项目(660509):广东石油 化工学院自然科学青年基金项目(650150), 备。通常为了提高腐蚀原图像质量,可以从内部 通信作者:张磊.E-mail:zhanglei@gdupt.edu.cn, 和外部两方面来进行实施):1)对图像采集的外DOI: 10.11992/tis.201712024 网络出版地址: http://kns.cnki.net/kcms/detail/23.1538.TP.20180507.0952.002.html 金属腐蚀区域图像增强算法研究 雷芳1 ,熊建斌2 ,张磊1 ,郭斯羽3 (1. 广东石油化工学院 计算机与电子信息学院,广东 茂名 525000; 2. 广东技术师范学院 自动化学院,广东 广 州 510665; 3. 湖南大学 电气与信息工程学院,湖南 长沙 410082) 摘 要:针对金属腐蚀区域图像中存在暗细节对比度不高、光照不均匀及颜色特征需保护的问题,提出一种在 HSI 模型下的多尺度细节自适应增强与同态滤波的增强算法。首先,对 RGB 腐蚀图像进行色彩空间变换,保 留其中的色调和饱和度分量不变,对亮度分量进行增强。然后,通过小波变换进行多尺度细节自适应增强,提 升细节对比度并作分块同态滤波,改善光照不均的影响,获得增强后的腐蚀图像。实验结果表明,所提方法增 加了腐蚀暗细节的对比度,提高了金属腐蚀区域图像的整体亮度并保证了色彩信息的不失真。 关键词:金属腐蚀图像;HSI 模型;多尺度;细节自适应增强;分块同态滤波 中图分类号:TP391 文献标志码:A 文章编号:1673−4785(2019)02−0385−08 中文引用格式:雷芳, 熊建斌, 张磊, 等. 金属腐蚀区域图像增强算法研究[J]. 智能系统学报, 2019, 14(2): 385–392. 英文引用格式:LEI Fang, XIONG Jianbin, ZHANG Lei, et al. Image enhancement algorithm in metal corrosion area[J]. CAAI transactions on intelligent systems, 2019, 14(2): 385–392. Image enhancement algorithm in metal corrosion area LEI Fang1 ,XIONG Jianbin2 ,ZHANG Lei1 ,GUO Siyu3 (1. College of Computer and Electronic Information, Guangdong University of Petrochemical Technology, Maoming 525000, China; 2. School of Automation, Guangdong Polytechnic Normal University, Guangzhou 510665, China; 3. School of Electrical and Information Engineering, Hu’nan University, Changsha 410082, China) Abstract: Considering the images of the metal corrosion areas, the dark details have low contrast, and the illumination is not uniform; meanwhile, the color characteristics need to be preserved. To solve these problems, an approach based on multi-scale detail-adaptive enhancement and homomorphic filtering is proposed on the basis of the HSI model. First, the RGB corrosion image was color-transformed, whereby the hue and saturation components were preserved, and the luminance component was enhanced. Then, wavelet transform was used to implement multi-scale detail-adaptive enhancement, increase the contrast of the detail, and apply block homomorphic filtering, so as to improve the impact of non-uniform illumination. Consequently, the corrosion image was enhanced. Experimental results show that the proposed method can increase the contrast of dark details and improve the overall brightness of the image of the metal corrosion area, ensuring that the color information is undistorted. Keywords: metal corrosion image; HSI model; multi-scale; detail-adaptive enhancement; block homomorphic filtering 金属材料受到环境因素的影响其腐蚀表面呈 现形貌纹理、灰度、颜色等不同特征。通过对这 些特征的分析可识别材料腐蚀的程度和腐蚀的类 型 [1-2]。在腐蚀无损检测的远程视觉检测 (remote vision inspection,RVI) 检测中,主要是使用视觉设 备来获取腐蚀信息。对于金属外部腐蚀,可以采 用摄像机或 CCD 相机获取腐蚀图像,设备内部腐 蚀图像的获取主要有管窥镜、自动爬行器等设 备。通常为了提高腐蚀原图像质量,可以从内部 和外部两方面来进行实施[3] :1) 对图像采集的外 收稿日期:2017−12−26. 网络出版日期:2018−05−09. 基金项目:国家自然科学基金项目 (61473331,61471167,61571147); 茂名市科技计划工业攻关项目 (660509);广东石油 化工学院自然科学青年基金项目 (650150). 通信作者:张磊. E-mail:zhanglei@gdupt.edu.cn. 第 14 卷第 2 期 智 能 系 统 学 报 Vol.14 No.2 2019 年 3 月 CAAI Transactions on Intelligent Systems Mar. 2019