摘要
高分辨率遥感影像中道路网的提取是智能地物提取和分析的重要方面,研究遥感图像的道路提取方法具有重要科学意义。本文以马萨诸州航空影像道路数据集为试验数据,设计一种基于端到端的全卷积神经网络模型,基于此分别对郊区道路网以及城区密集道路网进行试验分析,并与传统的监督分类算法进行比较。试验结果表明,基于卷积神经网络的深度学习算法可较高精度地从高分辨率遥感影像中提取道路信息,召回率与F1评分有着25%~47%左右的提升。
Road network extraction in high-resolution remote sensing images is an important aspect of intelligent ground feature extraction and analysis.It is of great scientific significance to study road extraction methods for remote sensing images.This paper uses the Massachusetts aerial image road data set as experimental data to design an end-to-end fully convolutional neural network model.Based on this,the suburban road network and the dense urban road network are tested and analyzed separately.Classification algorithms are compared.The experimental results show that the deep learning algorithm based on convolutional neural network can extract road information from high-resolution remote sensing images with high accuracy,and the recall rate and F1 score are improved by about 25% to 47%.
作者
王娟
谢跃辉
江丽芬
WANG Juan;XIE Yuehui;JIANG Lifen(Guangdong Province Geologica Surveying and Mapping Institute,Guangzhou Guangdong 510800,China;College of Geomatics and Geoinformation,Guilin University of Technology,Guilin Guangxi 541004,China)
出处
《北京测绘》
2020年第7期901-904,共4页
Beijing Surveying and Mapping
关键词
高分辨率遥感影像
全卷积神经网络
道路提取
high-resolution remote sensing image
full convolutional neural network
road extraction
