| 基于分区分层随机抽样的棉花种植面积遥感监测方法研究 |
| |
| 引用本文： | 胡永森,王力,施开分,周巍,饶华,王长耀.基于分区分层随机抽样的棉花种植面积遥感监测方法研究[J].河南农业科学,2016(10):155-160. |
| |
| 作者姓名： | 胡永森 王力 施开分 周巍 饶华 王长耀 |
| |
| 作者单位： | 1. 东华理工大学 测绘工程学院,江西 南昌330013;中国科学院遥感与数字地球研究所/遥感科学国家重点实验室,北京100101;2. 中国科学院遥感与数字地球研究所/遥感科学国家重点实验室,北京,100101;3. 国家统计局农村社会经济调查司,北京,100826;4. 国家统计局新疆建设兵团调查总队,新疆 乌鲁木齐,830000 |
| |
| 基金项目： | 国家统计局新疆棉花种植面积遥感调查项目，国家863 计划项目(2014AA06A511)，国家自然科学基金项目(41371358)，国家科技重大专项(14CNIC-032079-32-02) |
| |
| 摘 要： | 以新疆建设兵团为研究区域,依据当地作物种植结构,结合遥感技术,在借鉴前人对大区域作物面积遥感估算研究方法的基础上,采取分区分层随机抽样的方法对新疆建设兵团的棉花种植面积进行估算,并引入均值和变异系数等精度评价指标对该抽样方法行评估,最后根据外推估算方法,以遥感影像中的棉花种植面积统计结果为真值与抽样反推结果进行对比分析。结果表明,各组格网的棉花种植面积均值的估计量变异系数在试验中的变化幅度很小,抽样体系稳定,符合统计抽样原理对精度的高要求。反推后的各组棉花种植总面积精度均达到95%以上,反推质量效果较好。综合来看,上述空间抽样方法是可行的,尤其是对总体分区分层的划分,对后面的抽样及反推工作起到了很好的效果,可为政府对新疆建设兵团棉花种植面积的统计提供方法参考。
|
| 关 键 词： | 遥感 抽样 分层 比例分配 奈曼分配 |
Cotton Planting Area Estimation Method with Remote Sensing Based on Partitioned and Stratified Random Sampling |
| |
| Abstract: | According to local crop planting structure combined with remote sensing technology,with the Xinjiang Production and Construction Crops as the study area,the estimation of cotton growing area was done with the method of partitioned and stratified random sampling based on the previous studies research methods of remote sensing estimation for large area crop area. The mean and coefficient of variation were introduced to evaluate the effective of this sampling method. Finally, the data of cotton growing area in remote sensing images were compared with the results of the sampling estimation based on the extrapolation estimation method. The results showed that the variation range of the cotton planting area of mean estimator of coefficients of variation of each grid in the test was very small,and the sampling system was stable,which accorded with the high demand of the principle of statistical sampling to precision. The total area accuracy of cotton planting in each group reached more than 95% by extrapolation estimation method,and the extrapolation quality was high. Generally,this spatial sampling approach proposed in the article was feasible,especially the method of partitioned and stratified sampling improved the accuracy,which offered a reference for the government of the Xinjiang Production and Construction Crops cotton planting area extraction method. |
| |
| Keywords: | remote sensing samples stratification proportional sharing neyman allocation |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
