基于图像处理的铁路混凝土桥梁裂缝探测和信息获取
发布时间：2018-09-14 18:33
【摘要】：随着我国铁路里程数的不断增长和一些铁路线路服役时间的增加,铁路混凝土桥梁的安全检测任务越来越重。裂缝作为混凝土桥梁的严重病害是铁路混凝土桥梁安全检测的重要内容。目前铁路桥梁裂缝检测的形式主要以人工为主,它存在检测效率低、检测安全性低、劳动强度大、结果不准确等问题。近年来使用图像处理的方式进行裂缝检测得到了广泛关注。本文提出了一种基于图像处理的铁路混凝土桥梁裂缝的探测算法,并且实现了裂缝信息的自动获取,开发了一套用于铁路混凝土桥梁底面裂缝检测的检测设备,提高了桥梁裂缝检测的效率。 本文使用自行设计的检测设备进行桥梁底部的图像采集和相机标定,通过图像处理的技术对采集到的图片进行分析处理,有效的探测出图像中存在的裂缝,并且对这些裂缝实现了信息的自动获取,从而完成了对铁路混凝土桥梁的裂缝检测工作,设备可以有效的寻找出大于0.15mm的所有裂缝,满足了桥梁裂缝的最大限值。 在裂缝的探测过程中,本文使用了液体渗透的理论。在深入研究渗透模型的理论之后,提出了一种新的渗透算法,利用裂缝的形态学特征,可以在最大限度保留裂缝信息的情况下,有效的过滤掉所有的非裂缝干扰信息,得到一幅仅包含裂缝信息的二值图像,我们称之为裂缝地图。然后,通过对裂缝地图进行距离变换和细化处理等一系列形态学的变换,结合相机标定的结果,程序可以自动地求出裂缝的宽度。 最后,本文使用受试者工作特征分析技术(ROC)衡量了裂缝探测算法的性能,对自动测量得到的数据和人工测量的数据进行了误差分析,验证了本文提出算法的有效性。
[Abstract]:With the increasing number of railway mileage and the increasing service time of some railway lines, the safety detection task of railway concrete bridges becomes more and more serious. As a serious disease of concrete bridge, crack is an important content of railway concrete bridge safety detection. At present, the main form of railway bridge crack detection is manual. It has some problems, such as low detection efficiency, low detection security, large labor intensity and inaccurate results. In recent years, the use of image processing for crack detection has received extensive attention. This paper presents an algorithm for detecting cracks in railway concrete bridges based on image processing, and realizes the automatic acquisition of crack information, and develops a set of detecting equipment for detecting cracks on the bottom surface of railway concrete bridges. The efficiency of bridge crack detection is improved. In this paper, the self-designed detection equipment is used for image acquisition and camera calibration at the bottom of the bridge. The image collected is analyzed and processed by image processing technology, which effectively detects the cracks in the image. The information of these cracks is obtained automatically and the crack detection of railway concrete bridge is completed. The equipment can effectively find all cracks larger than 0.15mm and meet the maximum limit value of bridge cracks. In the process of crack detection, the theory of liquid permeation is used in this paper. After deeply studying the theory of permeation model, a new permeation algorithm is proposed. Using the morphological characteristics of cracks, we can filter out all non-crack interference information effectively under the condition of keeping crack information to the maximum extent. A binary image containing only crack information is obtained, which is called crack map. Then, through a series of morphological transformations such as distance transformation and thinning processing of the crack map, combined with the result of camera calibration, the program can automatically calculate the width of the crack. Finally, the performance of the crack detection algorithm is measured by using the operating feature analysis technique (ROC), and the error analysis of the data obtained from automatic measurement and that of manual measurement is carried out, which verifies the validity of the proposed algorithm.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U446;TP391.41
[Abstract]:With the increasing number of railway mileage and the increasing service time of some railway lines, the safety detection task of railway concrete bridges becomes more and more serious. As a serious disease of concrete bridge, crack is an important content of railway concrete bridge safety detection. At present, the main form of railway bridge crack detection is manual. It has some problems, such as low detection efficiency, low detection security, large labor intensity and inaccurate results. In recent years, the use of image processing for crack detection has received extensive attention. This paper presents an algorithm for detecting cracks in railway concrete bridges based on image processing, and realizes the automatic acquisition of crack information, and develops a set of detecting equipment for detecting cracks on the bottom surface of railway concrete bridges. The efficiency of bridge crack detection is improved. In this paper, the self-designed detection equipment is used for image acquisition and camera calibration at the bottom of the bridge. The image collected is analyzed and processed by image processing technology, which effectively detects the cracks in the image. The information of these cracks is obtained automatically and the crack detection of railway concrete bridge is completed. The equipment can effectively find all cracks larger than 0.15mm and meet the maximum limit value of bridge cracks. In the process of crack detection, the theory of liquid permeation is used in this paper. After deeply studying the theory of permeation model, a new permeation algorithm is proposed. Using the morphological characteristics of cracks, we can filter out all non-crack interference information effectively under the condition of keeping crack information to the maximum extent. A binary image containing only crack information is obtained, which is called crack map. Then, through a series of morphological transformations such as distance transformation and thinning processing of the crack map, combined with the result of camera calibration, the program can automatically calculate the width of the crack. Finally, the performance of the crack detection algorithm is measured by using the operating feature analysis technique (ROC), and the error analysis of the data obtained from automatic measurement and that of manual measurement is carried out, which verifies the validity of the proposed algorithm.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U446;TP391.41
【参考文献】
相关期刊论文 前8条
1 秦广胜;何对燕;商红林;;一种新的激光辅助钢板宽度测量系统的实现[J];电子设计工程;2009年10期
2 于海,石伟霞,赵志立;NUP型轴承内圈滚道宽度的精确测量[J];轴承;2005年11期
3 张维峰;刘萌;杨明慧;;基于数字图像处理的桥梁裂缝检测技术[J];现代交通技术;2008年05期
4 赵华;“三量”合理可采期的计算方法及其应用[J];科技情报开发与经济;2003年02期
5 万智萍;叶仕通;;基于OpenCV的双目立体视觉监控跟踪系统[J];科学技术与工程;2013年05期
6 邹振利;傅军;崔e，
本文编号：2243527
本文链接：https://www.wllwen.com/kejilunwen/jiaotonggongchenglunwen/2243527.html
