基于GEE平台的水稻田快速识别与提取方法  

Rapid identification and extraction method of rice paddies based on GEE platform

作　　者：覃楚仪 费腾[1] TAN Chuyi;FEI Teng(School of Resource and Environmental Sciences,Wuhan University,Wuhan,Hubei 430079,China)

机构地区：[1]武汉大学资源与环境科学学院,湖北武汉430079

出　　处：《北京测绘》2024年第6期827-833,共7页Beijing Surveying and Mapping

基　　金：国家自然科学基金(42271476)。

摘　　要：水稻是中国65%以上人口的主粮。及时准确地掌握水稻种植面积和空间分布,对于制定农业政策、保障粮食安全具有重要意义。本文提出了一种基于水稻种植模式的小型水体识别与阈值分割相结合的方法进行省域水稻田识别。该方法数据采用哨兵2号光学遥感影像,利用谷歌地球引擎(GEE)平台在广西壮族自治区进行了试验和精度验证。本文方法在水稻田识别方面表现出可靠、快速的性能,Kappa系数为0.76,总体精度为92%,生产者精度为70%,用户精度为95%。这种计算开销节约的方法实现了提取精度与计算效率的平衡,可有效应用在省域或更大区域的单季稻水稻田识别中。Rice is the staple food for more than 65% of Chinese people.A timely and accurate grasp of the cultivation area and spatial distribution of rice is of great significance for formulating agricultural policies and guaranteeing food security.In this paper,a method combining the identification of small water bodies based on rice cultivation patterns and threshold segmentation was proposed for identifying provincial rice paddies.The method adopted the data from optical remote sensing images of Sentinel-2,and tests and accuracy verification were carried out in the Guangxi Zhuang Autonomous Region using the Google Earth Engine(GEE) platform.The method proposed in this paper showed reliable and fast performance in identifying rice paddies,with a Kappa coefficient of 0.76,an overall accuracy of 92%,a producer accuracy of 70%,and a user accuracy of 95%.This approach saving computational overhead achieved a balance between extraction accuracy and computational efficiency and could be effectively applied in identifying rice paddies using a single-season rice cropping cycle in provincial or larger regions.

关 键 词：遥感指数 农业遥感 水稻田识别 哨兵2号 谷歌地球引擎(GEE) 

分 类 号：P237[天文地球—摄影测量与遥感]