基于无人机视觉的桥梁裂缝检测  

Bridge Crack Detection Based on Unmanned Aerial Vehicle Vision

作　　者：陶晓力 武建 杨坤 

机构地区：[1]南京航空航天大学计算机科学与技术学院,江苏南京211106 [2]中设设计集团股份有限公司,江苏南京210000 [3]连云港市公路管理处,江苏连云港222002

出　　处：《计算机技术与发展》2018年第3期174-177,共4页Computer Technology and Development

基　　金：国家自然科学基金(61375021);中央高校基本科研业务费专项资金;江苏省交通运输科技项目(2016T11);南京航空航天大学研究生创新基地(实验室)开放基金(kfjj20171608)

摘　　要：针对公路桥梁的养护和管理,传统的人工检测方法不仅工作危险性大、作业成本高,而且工作效率低、检测结果可靠性差。因此自动化的智能检测识别方式迫在眉睫。针对桥梁裂缝细小难获取的问题,采用无人机装配高倍变焦摄像头的方法来采集桥梁裂缝图像。采用最小值滤波、边缘检测等图像处理方法进行裂缝处理,通过链码跟踪的方法跟踪裂缝边缘得到裂缝周长面积等信息,并根据裂缝的线性特征进行特征的选择检测。针对裂缝图像检测过程中断裂的问题,根据裂缝连续、断点相近的特点,设计了一种基于裂缝线段最邻近端点的连接方法。通过实地对病害桥梁的考察获取了裂缝图像,对其进行检测处理,取得了良好的效果。In viewof maintenance and management for highway bridge,traditional manual testing has defects of high risk and cost,lowefficiency and poor reliability in checking.So the automated methods of intelligent detection and identification are imminent.In this paper,aiming at the problem that the bridge crack is difficult to obtain,we use the way of unmanned aerial vehicle(UAV) with high power zoom camera to collect the bridge crack image.The crack processing is carried out by means of image processing such as minimum filtering and edge detection.The information such as crack perimeter is obtained by tracking the edge of the crack by chain code tracking,and the crack is selected according to the linear characteristic.Aiming at the problem of fracture in the detection of crack image,we design a connection method based on the nearest neighbor end of the crack line according to the characteristics of continuity and breakage of the crack.Through the field investigation of the disease bridge,the crack images are got and processed with better results.

关 键 词：裂缝检测 无人机 沈俊边缘检测 裂缝连接 链码跟踪 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]