摘要
为探究天山北坡经济带2000—2020年农业大棚时空变化特征,在Google Earth Engine(GEE)平台上调用研究区1999―2020年多时相Landsat影像,结合同时期的GF-2、Quick Bird卫星的高分辨率影像在Landsat影像上提取农业大棚训练样本,构建了缨帽变换特征、光谱指数、纹理、地形等特征空间,在此基础上运用随机森林算法分别构建了基于Landsat 5和Landsat 8影像的农业大棚遥感提取模型,提取了研究区2000年、2005年、2010年、2015年和2020年5个时期的农业大棚空间分布图,并计算了研究区5个时期的农业大棚的景观指数和局部莫兰指数。结果表明:①构建的农业大棚遥感提取模型能准确地提取研究区各时期的农业大棚,总体精度在94.39%~99.41%之间,Kappa系数均达0.95以上;②2000年、2005年、2010年、2015年、2020年研究区农业大棚占地面积分别为7911.12 hm^(2)、16569.08 hm^(2)、22692.77 hm^(2)、22116.56 hm^(2)和19027.22 hm^(2),呈先增后降趋势;③21 a间乌鲁木齐县至昌吉(简称乌昌一带)农业大棚占地面积呈先增后减趋势,而吐鲁番市和精河县农业大棚占地面积持续增长;④研究区内农业大棚大斑块面积占比增多、破碎度下降、凝聚度提高,农业大棚发展趋势总体向好,正朝着集群化、规模化方向发展,但部分区域仍处于初级发展阶段,急需政策引导和调控。
In order to investigate the spatial and temporal characteristics of agricultural greenhouses in the north slope economic zone of Tianshan Mountains in the past 20 years,multi-temporal Landsat images of the study area from 1999 to 2020 were retrieved on the platform of Google Earth Engine(GEE),and the training samples of agricultural greenhouses were extracted from Landsat images by combining the high-resolution images of GF-2 and QuickBird satellites of the same period,and the feature space including tasseled cap transformation features,spectral index,texture,and topography was constructed by using the random forest algorithm.The feature space including tasseled cap transformation features,vegetation index,texture,topography,etc.was constructed,on the basis of which the remote sensing extraction model of agricultural greenhouses based on Landsat 5 and Landsat 8 images was constructed by using the random forest algorithm,and the spatial distributions of agricultural greenhouses in the study area in the five periods of 2000,2005,2010,2015 and 2020,respectively,were extracted,and the spatial distributions of agricultural greenhouses were calculated.and the landscape index and local Moran index of agricultural greenhouses in the study area were calculated.The results show that 1)The constructed remote sensing extraction model of agricultural greenhouses can accurately extract the agricultural greenhouses in the study area in each period,with the overall accuracy ranging from 94.39%to 99.41%,and the Kappa coefficients are all more than 0.95;2)The agricultural greenhouses area of the study area in 2000,2005,2010,2015 and 2020 were 7.91×103 hm^(2),16.57×103 hm^(2),22.69×103 hm^(2),22.12×103 hm^(2)and 19.03×103 hm^(2)respectively;showing an increasing and then decreasing trend;3)The agricultural greenhouse area of Urumqi County-Changji City showed an increasing and then decreasing trend,while the agricultural greenhouse area of Turpan City and Jinghe County showed a continuous increasing trend in the past 20 years;4)The proportion of large areas and the degree of cohesion of agricultural greenhouses increased,while the degree of fragmentation decreased,indicating that the development trend of agricultural greenhouses in the northern slope of Tianshan Mountains Economic Belt is moving in the direction of clustering and scale.However,some regions are still in the primary stage of development and urgently need policy guidance and regulation.
作者
陆笑舒
林芬芳
缪钱龙
吴春发
Lu Xiaoshu;Lin Fenfang;Miao Qianong;Wu Chunfa(School of Ecology and Applied Meteorology,Nanjing University of Information Science and Technology,Nanjing 210044,Jiangsu,China;School of Remote Sensing and Geomatics Engineering,Nanjing University of Information Science and Technology,Nanjing 210044,Jiangsu,China)
出处
《地理科学》
北大核心
2025年第7期1590-1600,共11页
Geographical Science
基金
第三次新疆综合考察项目(2021xjkk0903)资助。
关键词
农业大棚
随机森林算法
遥感提取
景观指数
agricultural greenhouses
Random Forest
remote sensing extraction
landscape indices
