摘要
定期开展码头水下结构的外观巡检是码头平稳运行的重要保障。通过分析现有的检测方法,发现获取图像质量不佳是采用视觉技术实施外观巡检的主要限制条件。结合相关检测作业的要求,创新提出将人造多幅欠曝光算法与颜色通道补偿方式进行融合,并通过试验验证使用图像增强算法的水下拍摄图像的视觉质量。结果表明,多曝光与颜色补偿可有效提升水下图像的对比度、恢复细节信息并进行颜色矫正,且运算效率高;图像增强算法可有效提升水下拍摄图像的视觉质量,帮助技术人员及时识别水下结构缺陷;试验中SIFT特征点匹配数量说明提出的算法有利于后续开展缺陷的自动识别及测量研究。
Regular inspection of the appearance of underwater structure is an important prerequisite for the smooth operation of a wharf.By the analysis of existing inspection methods,we find that the poor image quality is the main limitation of the application of the visual techniques in the appearance inspection of underwater structure.Combining with the requirement of inspecting underwater structure,we innovatively integrated the artificial multi-underexposure fusion strategy with the color channel compensation method,and verify the visual quality of the underwater image of the image enhancement method by test.The results show that the multi-exposure fusion and color compensation can effectively enhance contrast,recover details,and correct color-cast of the underwater image with high computation efficiency.The image enhancement method can improve the visual quality of underwater images,and help engineers recognize the existing deficiencies of the wharf in time.The number of the matched pairs of SIFT feature points indicates that the proposed algorithm is also beneficial to the subsequent research of automatic deficiency recognition and measurement.
作者
徐鲁强
陶冶
陈广桐
柏强
李翔
刘广会
XU Lu-qiang;TAO Ye;CHEN Guang-tong;BAI Qiang;LI Xiang;LIU Guang-hui(Centre of Technology,Liaoning Port Group Co.,Ltd.,Dalian 116004,China;Dalian Port Design&Research Institute Co.,Ltd.,Dalian 116023,China)
出处
《水运工程》
北大核心
2021年第6期42-49,共8页
Port & Waterway Engineering
基金
辽宁省交通科技计划项目(201902)。
关键词
水下结构外观检查
图像增强算法
视觉质量
运算效率
appearance inspection of underwater structure
image enhancement algorithm
visual quality
computational efficiency
