基于图像结构信息的隧道衬砌裂缝实时快速识别

裂缝是隧道衬砌最常见的病害之一,裂缝检测是后期隧道养护的前提。在图像采集过程中对隧道衬砌裂缝进行实时快速识别,对提高隧道病害检测效率具有重要的意义。针对传统的裂缝图像识别方法存在运行速度慢、效率低等缺点,提出基于图像结构信息的隧道衬砌裂缝实时快速识别方法。首先将图像结构信息用图像像素的灰度均值、标准差和结构元素掩码表示,随后采用不同方向的结构元素掩码对采集的当前图像进行卷积运算,并计算当前图像与样本图像之间的亮度信息和对比度信息,最后计算图像结构信息因子,可实时快速识别出衬砌裂缝,并通过接受者操作特征曲线估计最优参数和最佳阈值。现场实验结果表明,算法的识别时间远小于图像的采集和存储时间,识别速度快且准确性高,可显著提高隧道衬砌病害的检测效率,具有一定的工程应用价值。

Real-time Fast Recognition of Tunnel Lining Cracks Based on Image Structure Information

Crack is one of the most common diseases of tunnel lining, and crack detection is the premise of later tunnel maintenance. In the process of image acquisition, real-time fast recognition of tunnel lining cracks is of great significance to improve the efficiency of tunnel disease inspection. In view of the disadvantages of traditional crack image recognition methods with slow running speed and low efficiency, the real-time fast recognition method of tunnel lining cracks based on image structure information was proposed. Firstly, the image structure information is expressed by the gray mean value, standard deviation and structure element mask of image pixels. Then the current image is convoluted by means of different structural element masks, and luminance information and contrast information between the current image and sample images are calculated. Finally, the image structure information factor is calculated to recognize lining cracks in real time and quickly. The optimal parameters and threshold of the proposed algorithm are estimated using ROC curves. Experimental results show that the running time of the proposed algorithm is significantly less than the image acquisition and storage time with fast recognition speed and high accuracy. The study results can significantly improve the efficiency of tunnel lining diseases inspection and has certain engineering application value.