摘要
以邯郸市为研究对象,基于Google Earth Engine(GEE)平台与多源遥感数据,构建一种冬小麦训练样本自动生成与分类的方法。通过融合SNIC分割、物候特征筛选与随机森林算法,建立融合光谱、植被指数及纹理的多特征组合方案。结果显示,特征组合(3)(光谱+植被指数+纹理)的提取效果最优,其相对误差连续3年均为最低(2023年0.21%、2024年1.33%、2025年0.44%),总体精度和Kappa系数逐年提升。基于该方案生成了2023—2025年邯郸市冬小麦种植空间分布图,邯郸市冬小麦种植空间分布呈东部平原集中、西部山区较少的分异特征。长势监测显示,2025年邯郸市冬小麦在整个生育期内光照、温度、降水量、湿度条件匹配良好,NDVI增量以偏好为主,整体长势优于2023年和2024年。自动化样本生成方法在大范围作物分类中具备良好的适用性与稳定性。
Taking Handan City as the study area,a method for automatic generation of training samples and classification of winter wheat was constructed based on the Google Earth Engine(GEE)platform and multi-source remote sensing data.By integrating SNIC segmentation,phenological feature screening,and the random forest algorithm,a multi-feature combination scheme incorporat-ing spectra,vegetation indices,and texture was established.The results showed that feature combination③(spectra+vegetation indi-ces+texture)achieved the best extraction performance,with the lowest relative error for three consecutive years(0.21%in 2023,1.33%in 2024,and 0.44%in 2025),and the overall accuracy and Kappa coefficient improved annually.Based on this scheme,spa-tial distribution maps of winter wheat planting in Handan City from 2023 to 2025 were generated,revealing a distribution pattern con-centrated in the eastern plains and sparse in the western mountainous areas.Growth monitoring indicated that the light,temperature,precipitation,and humidity conditions during the entire growth period of winter wheat in Handan City in 2025 were well-matched,the NDVI increment was predominantly favorable,and the overall growth status was better than that in 2023 and 2024.The automatic gen-eration sample method demonstrated good applicability and stability in large-scale crop classification.
作者
李亚强
曹俊涛
常宇飞
孟成真
张珺
刀剑
赵春雷
权畅
LI Ya-qiang;CAO Jun-tao;CHANG Yu-fei;MENG Cheng-zhen;ZHANG Jun;DAO Jian;ZHAO Chun-lei;QUAN Chang(China Meteorological Administration Xiong’an Atmospheric Boundary Layer Key Laboratory,Xiong’an 071800,Hebei,China;Hebei Institute of Meteorological Sciences/Key Laboratory of Meteorology and Ecological Environment of Hebei,Hebei Meteorological Bureau,Shijiazhuang 050000,China;Handan Meteorological Bureau,Handan 056002,Hebei,China;College of Plant Protection,Yunnan Agricultural University,Kunming 650108,China)
出处
《湖北农业科学》
2025年第9期220-228,237,共10页
Hubei Agricultural Sciences
基金
中国气象局青年创新团队“高标准农田智慧气象保障技术”项目
风云卫星应用先行计划(2023)项目(FY-APP-ZX-2023.01)
河北省气象局青年基金项目(2025ky18
2023ky05)
海河流域气象科技创新项目(HHXM202507)。
