摘要
隧道衬砌的病害严重影响着行车和生命安全,裂缝是隧道衬砌最常见也是最严重的病害之一,需要对其进行定期检查和测量。基于数字图像技术设计隧道衬砌裂缝识别系统,对相关算法进行研究和优化。针对隧道环境内采集的图像容易曝光不足、光照不均且噪声严重的问题,对图像进行增强处理后构建基于双边滤波改进的去噪方法,可以在保护裂缝边缘细节的基础上滤除大量噪声。因有渗水、污渍、剥落等区域对隧道裂缝识别的影响,采用基于图像自适应分块下结合阈值和边缘信息的分割算法,有效地克服传统Otsu算法分割不准确的情况,得到完整裂缝的二值图像。采用相机尺寸标定,通过测量裂缝的像素尺寸转换得到其真实长度、宽度等指标,还可以进行裂缝等级评定。结合工程实例表明,所提出的算法对隧道衬砌裂缝的识别准确率达92%以上,验证了本文算法的有效性。
The lining of tunnels seriously affects driving and life safety. Cracks are one of the most common and serious distresses of tunnel lining, and they need to be regularly inspected and measured. In this paper, the crack identification system for tunnel lining was developed based on digital image technology, and the related algorithms were studied and optimized. This study aimed at conquering the problems that the images collected in the tunnel environment can be easily exposed and that the illumination is non-uniform and the noise is serious, the images were enhanced and the denoising method was constructed based on the improved bilateral filtering. The noise can be filtered out on the basis of the details of the crack boundaries. By accounting for the influences of seepage, stain and spalling on tunnel crack identification, the segmentation algorithm based on image adaptive segmentation and combined with threshold and edge information was used to effectively overcome the inaccurate segmentation by traditional Otsu algorithm and to obtain the complete binary images of tunnel cracks. By using the camera size calibration, the actual length and width of the cracks can be obtained by measuring the pixel size of the cracks, and the crack rating can also be performed. According to the applications in engineering examples, the accuracy of the proposed algorithm for identifying tunnel lining cracks is over 92%, which verifies the effectiveness of the proposed algorithm.
作者
唐钱龙
谭园
彭立敏
曹豪荣
TANG Qianlong;TAN Yuan;PENG Limin;CAO Haorong(School of Civil Engineering,Central South University,Changsha 410075,China;Jiangxi Vocational and Technical College of Communication,Nanchang 330013,China)
出处
《铁道科学与工程学报》
CAS
CSCD
北大核心
2019年第12期3041-3049,共9页
Journal of Railway Science and Engineering
基金
高铁联合基金资助项目(U1734208)
江西省教育厅科学技术研究资助项目(GJJ171292)
关键词
数字图像
隧道
衬砌
裂缝识别
digital image
tunnel
lining
crack identification