数字图像技术在露石混凝土路面检测中的应用研究

发布时间：2018-03-21 21:39

本文选题：露石水泥混凝土路面　切入点：构造深度　出处：《重庆交通大学》2015年硕士论文　论文类型：学位论文

【摘要】：露石水泥混凝土路面(EACCP)改善了普通水泥混凝土路面抗滑性能差、行车噪声大、易产生眩光等方面的缺陷,扩展了水泥混凝土路面的应用领域。相对于普通水泥混凝土路面,EACCP表面具有丰富的微观和宏观构造,表面处理相对较为复杂,其表面质量对路用性能有着决定性的作用。本文采用数字图像技术通过对露石水泥混凝土路面的表面图像的处理分析对其表面质量进行评价,在路面表面图像采集方法、图像信息与路面构造深度关系、露石颗粒分离与识别、EACCP表面分布均匀性评价指标与评价方法等方面,进行了深入的理论分析与应用研究。本文通过分析采用数字图像技术进行路面构造深度检测的原理,总结了图像采集过程中各种外界因素可能会对图像分析产生的影响,并以此为基础提出了规范化的图像采集方法,提高图像采集质量,为进一步的图像分析奠定了良好的基础;将采集得到的图像进行尺寸标定、裁切、平滑处理后,采用BP神经网络对EACCP表面图像信息和铺砂法实测路面构造深度的关系进行模拟,采用不同训练方法对不同规模的网络进行训练,并对得到的网络模型进行误差分析,得到符合一定精度要求的关系模型。针对EACCP图像中可能存在的露石颗粒与背景对比不明显、亮度不均的问题,采用分块-组合OTSU分割法,取得了较为理想的分割效果。采用数字图像技术提取EACCP表面露石颗粒信息,对其数量、面积以及静矩等指标进行计算分析,提出以被测表面露石颗粒数量、面积百分率的变异系数双指标对EACCP表面均匀性评价的方法,并借鉴相关规范中对公路技术状况的评价体系,提出路面均匀性状况指数PUI的概念及其计算方法,完善了EACCP表面均匀性评价体系。在MATLAB程序设计平台下对数字图像技术进行构造深度检测和露石表面均匀性检测评价的方法进行程序设计,利用GUI设计工具将其功能集成于一个易于理解和操作图形界面,并将其应用于重庆市涪陵区李渡镇EACCP试验路段表面质量检测和评价,验证了利用数字图像技术对EACCP表面构造深度测量、表面分布均匀性评价评价方法及程序界面的适应性。
[Abstract]:EAC CPC) improves the common cement concrete pavement with poor skid resistance, high driving noise and glare, etc. The application field of cement concrete pavement is expanded. Compared with ordinary cement concrete pavement, EACCP surface has abundant micro and macro structure, and surface treatment is relatively complex. The surface quality plays a decisive role in road performance. In this paper, digital image technology is used to evaluate the surface quality of exposed cement concrete pavement by processing and analyzing its surface image. The relationship between image information and pavement structure depth, the separation and identification of exposed rock particles, the evaluation index and evaluation method of surface distribution uniformity of EACCP, etc. This paper analyzes the principle of using digital image technology to detect the depth of pavement structure, and summarizes the influence of various external factors on image analysis in the process of image acquisition. On this basis, a standardized image acquisition method is proposed to improve the quality of image acquisition, which lays a good foundation for further image analysis. BP neural network is used to simulate the relationship between the surface image information of EACCP and the measured depth of pavement structure by sand laying method. Different training methods are used to train the network of different scales, and the error analysis of the obtained network model is carried out. The relation model which meets the requirement of certain precision is obtained. Aiming at the problem that the contrast between the exposed rock particles and the background is not obvious and the brightness is uneven in the EACCP image, a block-combinatorial OTSU segmentation method is adopted. The digital image technology is used to extract the information of exposed rock particles on the surface of EACCP, and the number, area and static moment of the information are calculated and analyzed, and the number of exposed rock particles on the measured surface is put forward. The method of evaluating the surface uniformity of EACCP by double index of variation coefficient of area percentage, and referring to the evaluation system of highway technical condition in relevant codes, the concept and calculation method of pavement uniform property condition index (PUI) are put forward. The evaluation system of surface uniformity of EACCP is perfected. The methods of structural depth detection and detection and evaluation of surface uniformity of exposed rock are designed on the platform of MATLAB programming. The function of GUI is integrated into a graphic interface which is easy to understand and operate, and it is applied to the testing and evaluation of surface quality of EACCP test section in Lidu Town, Fuling District of Chongqing City. The application of digital image technology to the depth measurement of EACCP surface structure, the evaluation of surface distribution uniformity and the adaptability of the program interface are verified.
【学位授予单位】：重庆交通大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U416.216;TP391.41

【参考文献】