深度学习遥感影像解译技术在园林绿化中的应用  

Application of Deep Learning Remote Sensing Image Interpretation Technology in Garden Greening

作　　者：张迎燕 曹奇 ZHANG Yingyan;CAO Qi(Changzhou Surveying and Mapping Institute,Changzhou,Jiangsu 213003,China;Changzhou Geospatial Information Intelligence Technology Center,Changzhou,Jiangsu 213003,China;Changzhou Vocational Institute of Textile and Garment,Changzhou,Jiangsu 213164,China)

机构地区：[1]常州市测绘院,江苏常州213003 [2]常州市地理信息智能技术中心,江苏常州213003 [3]常州纺织服装职业技术学院,江苏常州213164

出　　处：《自动化应用》2025年第4期63-65,69,共4页Automation Application

基　　金：2023年度江苏省高等学校自然科学研究项目资助(23KJB520001)。

摘　　要：以常州建成区城市绿化为研究对象,选取常州市区0.5m分辨率卫星遥感影像等作为数据源,构建可供训练的解译样本库,采用基于DeepLabv3+的深度学习遥感影像解译技术,创建一套园林绿化数据采集技术流程。通过这一流程,实现对“国家生态园林城市”创建指标的分析测算,指导辖市区绿化建设、养护和管理,辅助园林主管部门科学决策,从而推动城市的可持续发展。结果表明,深度学习遥感影像解译技术在园林绿化领域具有广阔的应用前景。Taking the urban greening in the built-up area of Changzhou as the research object,the 0.5 m resolution satellite remote sensing images of Changzhou urban area were selected as the data source to construct a sample library for training,and a set of garden greening data collection technology process was created by using the deep learning remote sensing image interpretation technology based on DeepLabv3+.Through this process,realize the analysis and calculation of the creation index of the"National Ecological Garden City",guide the construction,maintenance and management of greening in the urban area under its jurisdiction,assist the scientific decision-making of the competent department of gardening,and promote the sustainable development of the city.The results show that deep learning remote sensing image interpretation technology has a broad application prospect in the field of landscaping.

关 键 词：绿化数据采集 深度学习 遥感解译 语义分割 国家生态园林城市 

分 类 号：P237[天文地球—摄影测量与遥感]