摘要
提出一种路面破损图像实时检测的实现方法。对高速采集的路面图像去模糊后再进行裂纹特征分析,设计了一种基于邻域像素灰度值比较的特征快速提取方法,利用邻域相关及连贯性等知识智能去除虚假裂纹,从而提取真实裂纹信息。大量采集的图片的处理结果表明,该方法可实时快速检测路面破损信息,具有较强的鲁棒性和实际应用性。
A method for real-time detection of pavement distress is proposed.Firstly,the fissure characteristic of the fast acquired pavement image is analyzed after the blur is cleared, and the extracted method based on adjacent regions pixel greyvalue comparison is designed,the false cracks are deleted by using neighbourhood relativity and coherence of crack features, thus the true fissures are detected.Quantities of images are processed and the results show that the method can detect the pavement distress information quickly, and has robustness and actual availability.
出处
《电子设计工程》
2009年第6期36-37,40,共3页
Electronic Design Engineering
基金
国家自然科学基金资助项目(60572040)
关键词
路面破损
裂纹特征
图像分析
特征提取
pavement distress
fissure features
image analysis
feature extraction
