基于深度学习理论的山地城市水土保持卫星影像变化图斑提取实践  

Spot Extraction from Satellite Images of Soil and Water Conservation in Mountainous City Based on Deep Learning

作　　者：舒文强 蒋光毅 郭宏忠 张志兰 文力 SHU Wenqiang;JIANG Guangyi;GUO Hongzhong(Chongqing Geomatics and Remote Sensing Center,Chongqing 401147,China;Chongqing Soil and Water Conservation Monitoring Station,Chongqing 401147,China;Chongqing Municipal Water Resources Bureau,Chongqing 401147,China)

机构地区：[1]重庆市地理信息和遥感应用中心,重庆401147 [2]重庆市水土保持监测总站,重庆401147 [3]重庆市水利局,重庆401147

出　　处：《中国水土保持》2022年第5期26-29,I0005,共5页Soil and Water Conservation in China

摘　　要：随着卫星遥感技术的不断发展,遥感监测已成为及时精准发现人为水土流失违法违规行为的有效手段。重庆高山环绕、丘陵广布且常年云雾,影像采集困难,图斑破碎,是典型的山地城市,传统的遥感人工目视解译需要耗费大量的人力物力,难以满足大区域山地城市的变化图斑提取需求。为提升重庆市水土保持遥感监测效率,将深度学习理论应用到对不同时期影像的变化检测分析中,采用基于语义信息的变化图斑提取和基于端到端深度网络的变化图斑提取方法,能够自动发现疑似水土保持扰动变化图斑,大大提高了水土保持监测监管的工作效率。Along with the continuous development of satellite remote sensing technology,remote sensing monitoring has become an effective means for timely and accurate detection of human-induced violations of soil and water loss. Chongqing is a typical mountainous city,surrounded by high mountains,widely covered with hills and cloudy and foggy all year round. It is difficult to collect remote sensing images and the land use parcels are generally broken. Traditional remote sensing manual visual interpretation requires a lot of manpower and material resources and it is difficult to meet the demand for extraction of changing maps in large-scale mountainous cities. In order to improve the efficiency of remote sensing monitoring of soil and water conservation of Chongqing,the paper applied the Deep Learning theory to the change detection and analysis for the images in different periods by extracted the change pattern based on the methods of semantic information and the change pattern based on end-to-end depth network. It could automatically find the suspected change pattern of soil and water conservation disturbance and had greatly improved the work efficiency of soil and water conservation monitoring and supervision.

关 键 词：水土保持监测 遥感 变化检测 深度学习 重庆 

分 类 号：S157[农业科学—土壤学]