山岭隧道结构裂缝特征自动化提取算法研究

Automatic Extraction Algorithm of Mountain Tunnel Structural Crack Feature

基于机器视觉的缺陷检测技术日益成为隧道衬砌结构病害检测的研究重点,结构裂缝的自动化检测是其中的一个重要分支。针对公路隧道衬砌结构裂缝图像处理中的预处理环节,本文提出灰度变换增强算法结合频域滤波的方法,有效增强目标区域,平滑噪声区域,在进行灰度阈值分割图像的基础上,根据裂缝自身的形状和走向特征,利用面积、凸性、紧密度三大特征参数对裂缝区域进行提取。提取结果表明,该方法能准确提取图像中的裂缝区域。在此基础上,计算裂缝区域的几何参数信息,并结合现行公路隧道的相关规范要求,为评估隧道衬砌结构的安全性提供依据。

Defect detection technique based on vision detection has become a major researching of tunnel lining structural damage detection,of which the automated detection of structural cracks is an important branch.For the preprocessing in image processing of the lining cracks of road tunnel,this article presents an algorithm to combine the gray-scale transformation algorithms and the frequency domain filtering algorithm,which effectively enhance the target region and smooth the noise region.Based on image segmentation by means of gray-level threshold,and according to the shape feature and direction feature of the crack itself,this article extracts the crack region by using area,convexity,compactness parameters.The results show that crack region can be extracted from image correctly using the proposed algorithm.On this basis,we can calculate the geometric parameters of crack regions,which provides evidence for evaluating structural security in combination with the related requirements of highway tunnel.