摘要
为了识别路面的三维裂缝,根据路面病害的特征提取与检测要求,针对路面三维裂缝数据的特点,结合数字图像处理滤波技术,提出了一种基于平均值法曲线拟合的双相扫描检测路面裂缝识别方法。识别过程中先采用双相标准差均值滤波算法与基于组合结构元素的级联形态学滤波算法相结合的方法对数据点进行去噪处理,再通过基于平均值法曲线拟合的双相扫描检测方法对裂缝进行识别;即对路面三维数据矩阵分别进行水平单相和垂直单相扫描检测得到2幅裂缝图像,然后对这2幅裂缝图像进行叠加处理,再进行区域生长,便可得到双相扫描检测识别的裂缝图像。研究结果表明:采用双相扫描检测方法,可以很好地识别裂缝;采用准确率曲线对该算法进行评估,得出该算法性能稳定且准确率高;该算法对路面养护管理决策的制定和路表三维形态的重构起到了重要的指导作用。
In order to identify 3D pavement cracks,according to the pavement distress feature extraction and detection requirements,combined with digital image filtering processing technology,a 3D pavement crack identification method based on dual-phase scanning detection was put forward.A specific method was used to deal with the noise data points.This specific method combined the dual-phase standard deviation average filtering with the cascade morphological filtering algorithm based on composite structural elements.Then cracks could be identified through the dual-phase scanning detection based on the average method of curve fitting.One crack image was obtained based on horizontal single-phase scanning detection for pavement 3D data matrix.Another crack image was obtained based on vertical single-phase scanning detection for pavement 3D data matrix.The superposition for two crack images and region growing operation were processed.Crack image was then identified by dual-phase scanning detection.The results show that the method can identify cracks better;the algorithm given in the paper has good stability and high accuracy by accuracy rate curve assessment analysis.The paper provides an important guidance for the decision-making in pavement maintenance management and pavement 3D morphological reconstruction.
出处
《中国公路学报》
EI
CAS
CSCD
北大核心
2015年第2期26-32,共7页
China Journal of Highway and Transport
基金
交通运输部西部交通建设科技项目(2011 318 490 750)
陕西省青年科技新星项目(2013kjxx-93)
关键词
道路工程
裂缝识别
双相扫描
路面三维数据
区域生长
road engineering
crack identification
dual-phase scanning
pavement 3D data
re-gion growing
