基于卷积神经网络的光学遥感影像道路提取方法研究进展  

Research Progress of Road Extraction Method for Optical Remote Sensing Images Based on Convolutional Neural Network

作　　者：林雨准 刘智 王淑香 芮杰[1] 金飞[1] Lin Yuzhun;Liu Zhi;Wang Shuxiang;Rui Jie;Jin Fei(Institute of Geospatial Information,PLA Strategic Support Force Information Engineering University,Zhengzhou 450001,China)

机构地区：[1]战略支援部队信息工程大学地理空间信息学院,郑州450001

出　　处：《吉林大学学报（地球科学版）》2024年第3期1068-1080,共13页Journal of Jilin University：Earth Science Edition

基　　金：国家自然科学基金项目(42201443)。

摘　　要：随着光学遥感影像空间分辨率的提升和获取渠道的丰富,利用光学遥感影像实现地物智能解译已成为高效的技术路径。由于卷积神经网络(convolutional neural networks,CNN)强大的特征提取能力以及道路信息在多个领域的应用需求,基于CNN的道路提取方法成为了当前的研究热点。鉴于此,本文根据近年来的相关研究文献,对基于CNN的道路提取方法从基于形状特征的改进、基于连通性的改进、基于多尺度特征的改进和基于提取策略的改进四个方面进行归纳总结,然后描述典型道路遮挡案例,并利用经典CNN从样本标签的局限性层面对当前的技术难点进行分析与验证,最后从多源数据协同、样本库建设、弱监督模型和域适应学习四个方面对遥感影像道路提取的发展趋势进行评估和展望。With the improvement of spatial resolution of optical remote sensing images and the enrichment of acquisition channels,optical remote sensing images has become an efficient technological method to achieve intelligent interpretation of land features.Due to the powerful feature extraction ability of convolutional neural networks(CNN)and the demand of road information in many fields,road extraction methods based on CNN have become a current research hotspot.In view of this,this paper summarizes the road extraction method based on CNN from four aspects:Improvement of shape features,improvement of connectivity,improvement of multi-scale features and improvement of extraction strategy according to the relevant research literature in recent years.Then,we describe typical road occlusion cases and use classical CNNs to analyze and validate the current technical difficulties at the level of limitations of sample labels.Finally,the development trends of road extraction from remote sensing images are outlooked from four aspects,namely,multi-source data synergy,sample library construction,weakly supervised modeling and domain-adaptive learning.

关 键 词：卷积神经网络 光学 遥感影像 道路提取 智能解译 

分 类 号：P237[天文地球—摄影测量与遥感]