摘要
随着社会经济的快速发展,道路的交通量也呈现日益增长的趋势,基于遥感影像的路面健康检测和识别工作尤为重要。试验基于GF-2遥感影像数据,构建光谱指数和灰度共生矩阵,增加路面材质类型识别的波段因子,结合BP神经网络筛选最佳波段组合,利用SVM分类方法实现道路路面材质类型的识别。结果表明,基于最佳波段组合的道路路面材质识别结果均优于基于原始影像和主成分分析影像的识别结果,识别精度满足要求。该方法可快速、高效识别大区域的道路路面材质类型,为道路的路面健康检查和评价提供数据支撑。
With the rapid development of socio-economy,the traffic volume is showing an increasing trend,and road surface detection and identification work is particularly important.In this paper,based on GF-2 remote sensing image data,a spectral index and a gray level co-occurrence matrix are constructed,and a band factor for pavement material type recognition is added.The optimal waveband combination is selected in combination with BP neural network,and the road pavement material type recognition is realized using SVM classification method.The results show that the road surface material recognition results based on the best combination of bands are better than the recognition results based on the original image and the principal component analysis image,and the recognition accuracy meets the requirements.This method can quickly and efficiently identify the types of road pavement materials in large areas,and provide data support for road pavement health inspection and evaluation.
作者
肖国峰
张蕴灵
杨璇
郭丽丽
XIAO Guo-feng;ZHANG Yun-ling;YANG Xuan;GUO Li-li(China Highway Engineering Consultants Corporation,Beijing 100097,China;Research and Development Center of Transport Industry of Spatial Information Application and Disaster Prevention and Mitigation Technology,Beijing 100097,China;China Highway Engineering Consultants Corporation Data Co.Ltd.,Beijing 100097,China;Gansu Institute of Nature Resources Planning and Research,Lanzhou 730000,China)
出处
《公路》
北大核心
2020年第10期18-25,共8页
Highway
基金
高分辨率对地观测系统重大专项(民用部分),项目编号07-Y30B03-9001-19/21。
关键词
路面材质类型
GF-2遥感影像
灰度共生矩阵
光谱指数
BP神经网络
pavement material type
GF-2 remote sensing image
gray level co-occurrence matrix
spectral index
BP neural network
