摘要
本文提出了一种路面破损图像测量方法,即将路面破损图像等分为64×64象素的子块,选取路面破损子块图像的灰度方差值描述子块图像特征,采用了神经网络分类器识别路面破损子块图像模式,并在路面破损子块模式识别的基础上,确定路面破损面积的测量方法。在测量裂缝宽度时,首先确定破损最严重的裂缝子块区域,然后对该子块进行中值滤波、二值化等图像处理,最后通过计数的方法确定裂缝的宽度。
An image measuring method of pavement distress is put forward in this paper.The pavement surface image is divided into subimages of 64×64 pixels and the intensity variances are used to represent the subimage feature.Based on the result of the pavement subimages model recognition,the pavement distress area measuring method is determined.When we measure the cracking width,firstly find out the most severe cracking subimages,then carry on median filter and binarization processing,finally measure cracking width through counting the pixels of subimage.
出处
《公路交通科技》
CAS
CSCD
北大核心
2004年第7期12-16,共5页
Journal of Highway and Transportation Research and Development
基金
教育部博士点基金资助项目(20000185070)
关键词
路面破损
图像测量
图像处理
中值滤波
Pavement surface distress
Image measuring
Image processing
Median filter
