【作者】 柳钰；

【导师】 陈静；

【作者基本信息】 武汉理工大学 ， 模式识别与智能系统， 2007， 硕士

【摘要】 随着我国沥青公路建设的飞速发展，沥青公路在公路总里程中所占的比例越来越大，沥青公路建成后的管理与养护工作也需要快速的跟进；由于沥青公路使用年限为10年，我国早期的沥青公路已出现破损情况。车辙是沥青公路主要病害之一，破坏路面，危害行车安全。因此，如何对庞大的沥青公路资源进行快速、客观、准确的检测，对我国沥青公路工程施工质量评价及养护管理具有重要意义。本文主要对沥青路面车辙深度检测系统进行研究。针对目前沥青路面车辙检测系统的关键问题，根据系统的主要功能和技术指标，在国内外先进技术的基础之上，选择合适的检测方法，提出解决关键问题的方案。本文主要完成以下3方面工作：(1)研究国内外检测技术现状，结合检测系统的主要功能和技术指标，综合国内外的先进技术，选择合适的检测方案，采用基于线结构光的主动视觉法为检测方案，结合图像处理技术，实现对沥青路面进行无接触、无破损、高速度和高精度检测。；(2)针对实际的路面检测环境，选择合适的CCD摄像机及图像采集卡，并与特定的工控机平台相结合，组成图像分析系统的主要硬件环境，选取有效的路面图像算法完成路面图像的处理工作；该算法的主要步骤分两步：1)分别采用阈值分割和灰度邻域属性法消除白噪声及孤点噪声；2)使用B样条拟合重心法完成图像的细化工作。(3)对沥青路面的车辙评价进行了分析。从行车安全的角度，对现有车辙各评价指标进行了讨论，并提出了新的评价指标设想，对新指标进行了初步探讨，各评价指标的确可以描述车辙的不同特点。更多还原

【Abstract】 With the rapid development of national asphalt highway construction, the occupancy which asphalt highway possess at highway overall mileage is larger and larger. It is necessary to develop the maintained job immediately after asphalt highway has constructed. Because of the asphalt highway can only serve 10 years, the asphalt highway which has been constructed early has damaged in our country. Rut is one of the primary damage of the asphalt highway, it imperil the pavement and compromise the security of driving. Therefore, how to make the precise, impersonal and fast detection with enormous asphalt highway resources is supposed to be significant to the appraisement on the construction and maintenance on national asphalt highway.In this dissertation, focuses on the depth of rut in asphalt pavement surface detection system. In view of existence key problem of the depth of rut in asphalt pavement surface detection system and main function and the technical specification, studying the domestic and foreign advance technology, selecting the method of detection, proposing the solution to key problem.In this dissertation has solved three questions.Firstly, investigating present situation of the domestic and foreign technology, incorporate with main function and the technology, selecting the method of solution. It takes initiative visual technology which based on linear laser as the scheme of the detection, makes the detection without touch, damage and in rapidly, high precision.Secondly, in view of actual environment of detection, selecting proper CCD camera and image collection card, combining it with industry computer, composing main hardware of image analyze system, designing effective image algorithm for the system. The algorithm includes two major phases: 1) removing white noise and acnode noise by threshold segmentation and grayscale neighborhood property respectively; 2) using B-Spline based barycenter algorithm to finish the pavement image thinning.Finally, analyze appraise parameters of the rut. From aspect of driving security, discussing the present appraise parameters, supposing new parameters to describe different characteristic of rut.更多还原