摘要
提出了一种减少沥青路面破损图像识别计算量的图像分割方法。将路面图像等分为64×64像素的子块图像,并用灰度方差值描述子块图像特征。同时,设计了基于BP人工神经网络的子块图像模式分类器,利用子块图像模式分类结果所组成的矩阵作为路面破损图像分割结果。实验表明,该方法可较好地分割路面破损图像。
A method of asphalt pavement surface distress image segmentation is put forward,and it may reduce calculation of pavement surface distress image classify. The pavement surface image is divided into 64×64 pixel subimages,the intensity variances are used to represent the subimage feature.Meanwhile,the subimage pattern classifier is designed based on BP artificial neural network,all of the subimage pattern classifying results are arrayed a matrix, and the pavement surface distress image feature is represented by this matrix.The experiment shows that this matrix can represent the pavement surface distress image preferably.
出处
《中国公路学报》
EI
CAS
CSCD
北大核心
2003年第3期11-14,共4页
China Journal of Highway and Transport
基金
教育部博士点基金项目(2000018507)
关键词
道路工程
路面破损
图像识别
神经网络
模式分类器
road engineering
pavement surface distress
image recognition
neural network
pattern classifier