摘要
结合我国公路养护管理需要,研究了路面图像处理和损坏模式识别技术,提出了沥青路面损坏模式识别方法,解决了多种损坏类型并存图像的损坏模式识别关键技术。方法的基本思路是:依次分析、诊断坑槽、龟裂、块裂、纵裂和横裂等损坏,然后确定各类损坏的区域、量化损坏面积和长度,计算路面损坏率。实例验证表明,文章所介绍的路面损坏模式识别软件能有效地识别路面的坑槽、龟裂、不规则裂缝、纵裂和横裂,识别率达到95%以上。
A technology for pavement distress auto-identification for highway maintenance and management was studied. A method dealing with the identification and classification of multi-typed distresses of a single image was presented. The idea of the method is checking each type of distresses of the given pavement image and identifying its major characteristics such as cracking area, length, width, position and ratio of distressed area. Examples verify that the software can efficiently identify the major characteristics such as cracking area, length, width, position and ratio of distressed area, and the identification rate can be as high as 95% and more.
出处
《交通与计算机》
2007年第4期92-95,共4页
Computer and Communications
关键词
路面损坏
损坏类型
图像处理
模式识别
pavement distress
type of distress
image processing, pattern recognition