摘要
针对隧道内采集的衬砌图像普遍存在光照不均和混凝土坑洼不平的噪声的现象,导致在检测不同宽窄裂缝时准确率不高的问题,本文提出一种基于屋脊特征分析的裂缝检测算法,把在图像二维空间中的曲线检测问题转换到一维空间中的极值点检测问题。首先,通过检测极值点提取所有屋脊状特征;然后分析裂缝和噪声的区别,根据屋脊高度直方图检测裂缝;最后用数学形态学连接邻近断续线条完成裂缝的提取。该算法在辽宁省唐岭山隧道现场采集的实际衬砌图库中进行了实验验证,并与马尔科夫图像分割和小波变换边缘检测方法对比,裂缝检出率和噪声去除率都取得了更好的效果。
In view of the common phenomenon of uneven illumination and uneven concrete noise in the lining images collected in the tunnel,the accuracy rate is not high when detecting different width and narrow cracks.This paper proposed a crack detection algorithm based on ridge feature analysis.The curve detection problem in the two-dimensional space of the image was transformed into the extreme value point detection problem in the one-dimensional space.First,all ridge features by detecting extreme points were extracted.Then the difference between cracks and noise were analyzed,and cracks according to the roof height histogram were detected.Finally,mathematical morphology was used to connect adjacent segment lines to complete cracks extraction.The algorithm was experimentally verified in the actual lining gallery captured at the Tanglingshan tunnel in Liaoning province,and compared with Markov image segmentation and wavelet transform edge detection methods,the crack detection rate and noise removal rate achieve better results.
作者
薛丹
苑玮琦
XUE Dan;YUAN Wei-qi(Computer Vision Group,Shenyang University of Technology,Shenyang 110870,China)
出处
《仪表技术与传感器》
CSCD
北大核心
2022年第1期109-113,共5页
Instrument Technique and Sensor
关键词
隧道衬砌
裂缝检测
屋脊
光照不均
极值
tunnel lining
crack detection
ridge
uneven illumination
extreme value
