有监督学习下的市域铁路隧道结构裂缝边缘识别方法

A Detecting Method for Structure Crack Edges of Municipal Railway Tunnels under Supervised Learning

裂缝病害的定期巡检工作是保证市域铁路隧道结构正常运营的关键。针对市域铁路隧道结构表观裂缝的有效识别问题,提出了有监督学习下的市域铁路隧道结构裂缝边缘识别方法。首先,利用消费级数码相机获取了市域铁路隧道结构表观状态的图像数据;然后,基于Canny算子的初步边缘目标检测结果,构造了边缘目标的几何特征矢量,并在此基础上提出了基于几何特征数据的裂缝边缘识别方法;最后,采用实际市域铁路隧道结构的表观图像数据验证了所提出方法的有效性。

Regular inspection of crack diseases was the key to ensure the normal operation of metro tunnel structures.To address this issue,a method for crack detection of metro tunnel structures under supervised learning was proposed herein.Firstly,the images of the metro tunnel structures were obtained by a consumer digital camera.Secondly,the geometric feature vectors of the edge targets were constructed based on the preliminary detection results by Canny operator,and then the crack edges of metro tunnel structures were detected by using the geometric feature vectors of edge targets.Finally,the effectiveness of the proposed method was verified by the image data of the metro tunnel structures.