基于MMS图像的路面裂缝检测分析

发布时间：2018-06-24 05:49

本文选题：移动测量系统 + 路面裂缝检测　；参考：《北京建筑大学》2017年硕士论文

【摘要】：随着公路建设的日益完善,相关公路的检查与养护工作已经成为公路管理中的重要任务。针对当前人工检测道路裂缝方式存在耗时长、成本高和具有危险性等问题,文章研究一种基于数字图像处理的方法检测道路裂缝,可以较为快速、准确地识别路面裂缝信息。实验检测数据来源于移动测量系统(Mobile Mapping System,简称MMS)采集到的路面影像——以常规车速在公路行驶的同时利用车顶架设的摄像机采集公路路面的可量测立体影像信息。因此,重点就基于MMS图像的路面裂缝检测与分类算法进行研究。由于MMS图像存在硬件设备、光照不均和路面阴影等因素的影响,增加了图像处理的复杂度。通过对比分析确定较为理想的图像预处理步骤与方法:具体为采用加权平均值法进行图像灰度化;以降低计算量为目的的改进型双边滤波图像去噪算法;基于图像饱和度和光照比例因子去除图像阴影干扰;以及保证裂缝细节信息不被破坏的粗糙集理论算法增强图像。其次,分析目前两类经典的图像分割算法。针对单一型图像分割算法在检测裂缝中均存在一定的局限性,研究采用一种融合阈值法与数学形态学法的图像分割方法:即针对每幅MMS图像特点利用最大类间方差求出各图像的自适应阈值,并结合以多种方向的结构元素为基础的数学形态学算法实现道路裂缝的分割。此外,采用基于形态学生长的算法对分割结果中断裂的裂缝进行边缘连接,填补裂缝中未被提取到的灰度较低区域。最后,在路面裂缝检测结果的基础上,对路面裂缝特征进行描述和类别界定,并根据相关标准对路面损坏状况进行分析与评价。根据研究和实验验证,针对MMS图像检测道路裂缝的方法在保证检测准确性的同时,也具有较高的路面裂缝检测速度(通过优化算法),有助于快速有效地发现公路路面裂缝,对进一步实现道路裂缝检测自动化和实时化有着积极的推动作用,提升公路养护工作的智能化水平。
[Abstract]:With the improvement of highway construction, the inspection and maintenance of highway has become an important task in highway management. Aiming at the problems of long time, high cost and dangerous in manual detection of road cracks, this paper studies a method based on digital image processing to detect road cracks, which can identify the information of road cracks more quickly and accurately. The experimental data come from the road image collected by the Mobile Mapping system (MMS), which uses the camera mounted on the top of the vehicle to collect the stereo image of the road surface with the normal speed while driving on the road. Therefore, the algorithm of pavement crack detection and classification based on MMS image is studied. The complexity of MMS image processing is increased because of the influence of hardware equipment, uneven illumination and road surface shadow. Through comparative analysis, the ideal image preprocessing steps and methods are determined: the weighted average method is used to grayscale the image, and the improved bilateral filtering image denoising algorithm is designed to reduce the computation cost. Based on image saturation and illumination ratio factor, shadow interference is removed, and rough set theory algorithm is used to enhance the image, which ensures that the crack details are not destroyed. Secondly, two classical image segmentation algorithms are analyzed. Aiming at the limitation of single image segmentation algorithm in detecting cracks, In this paper, an image segmentation method based on fusion threshold method and mathematical morphology method is proposed. According to the characteristics of each MMS image, the adaptive threshold of each image is obtained by using the maximum inter-class variance. Combined with the mathematical morphology algorithm based on the structural elements in many directions, the road crack segmentation is realized. In addition, an algorithm based on morphological growth is used to connect the broken cracks in the segmentation results to fill the lower gray level areas that have not been extracted from the cracks. Finally, based on the detection results of pavement cracks, the characteristics of pavement cracks are described and classified, and the pavement damage status is analyzed and evaluated according to the relevant standards. According to the research and experimental verification, the method of road crack detection based on MMS image not only ensures the accuracy of the detection, but also has a higher detection speed (through the optimization algorithm), which is helpful to quickly and effectively find the road surface crack. It plays an active role in realizing the automation and realtime of road crack detection and improving the intelligent level of highway maintenance.
【学位授予单位】：北京建筑大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U418.6;TP391.41

【参考文献】