排水管道闭路电视视频智能检测技术的现状与挑战  

Current status and challenges of CCTV video intelligent detection technology for urban pipelines

作　　者：王大成 陈国强 秦军[2] 李国桥 Wang Dacheng;Chen Guoqiang;Qin Jun;Li Guoqiao(Guangdong Huiyu Intelligent Survey Technology Co.,Ltd.,Guangzhou 510665,China;Anhui University,Hefei 230601,China)

机构地区：[1]广东绘宇智能勘测科技有限公司,广州510665 [2]安徽大学,合肥230601

出　　处：《工程勘察》2022年第3期52-56,共5页Geotechnical Investigation & Surveying

摘　　要：城市排水管道缺陷检测是及时发现管道安全隐患的重要保障,为管道养护、修复提供准确的科学依据。人工管道缺陷检测方法费时费力,主观误差大,与其相比,管道闭路电视检测系统(Closed Circuit Television, CCTV)管道内窥检测方法自动化程度更高,但仍处于发展的初级阶段。本文首先分析人工检测方法与传统机器学习CCTV视频管道缺陷检测方法的优缺点,然后着重对深度学习CCTV视频管道缺陷智能检测方法进行介绍,分析该方法具有的优势与现阶段面临的挑战。最后梳理并展望深度学习智能化管道缺陷检测技术未来发展趋势。Urban drainage pipeline defect detection is an important guarantee for timely finding pipeline safety problems, and providing accurate data for pipeline maintenance and repair. Artificial pipeline defect detection method is time-consuming and laborious, with large subjective error. Compared with the artificial pipeline defect detection method, the CCTV video intelligent detection method has a higher degree of automation, but it is still in the initial stage of development. This paper firstly analyzes the advantages and disadvantages of the manual and traditional machine learning CCTV video method of pipeline defect detection, then emphatically introduces the intelligent detection method of deep learning CCTV, and analyzes its advantages and challenges at the present stage. Finally, we summarize and look forward to the future development trend of deep learning intelligent pipeline defect detection technology.

关 键 词：管道缺陷 目标识别 深度学习 智能化 

分 类 号：P258[天文地球—测绘科学与技术]