基于分层抽样的水稻种植面积回归估计

发布时间：2018-04-29 00:36

  本文选题：水稻种植面积 + 中分辨率数据　； 参考：《江苏农业科学》2017年16期

【摘要】：随着农业统计需求和遥感技术的发展,中国农业部即将使用中分辨率遥感数据开展中国主要农作物的全覆盖遥感调查,如何修正全覆盖调查结果使其更接近农作物面积真值是需要研究的问题。以水稻为例,在区域尺度上利用基于样本无人机数据解译得到的水稻种植面积修正基于中分辨率数据解译得到的水稻种植面积,分别比较分层分别回归估计和分层合并回归估计,利用基于全覆盖无人机数据解译得到的水稻种植面积验证回归估计值。在给定5%误差和95%置信水平下,抽样比为0.36%(最小样本量公式换算)的抽样结果表明:(1)2种方法的估计精度都大于95%;(2)合并比估计的精度好于分别比估计。
[Abstract]:With the development of agricultural statistics and remote sensing technology, the Ministry of Agriculture of China will soon use medium-resolution remote sensing data to carry out a full-coverage remote sensing survey of major crops in China. It is necessary to study how to correct the results of full coverage survey and make it closer to the true value of crop area. Taking rice as an example, using sample UAV data to interpret rice planting area correction based on mid-resolution data, rice planting area was obtained at regional scale. The hierarchical regression estimation and the hierarchical combined regression estimation were compared respectively. The rice planting area based on the full cover UAV data interpretation was used to verify the regression estimate. Under the given 5% error and 95% confidence level, the sampling results with a sampling ratio of 0.36 (conversion of the minimum sample size formula) show that the accuracy of the combined ratio estimation of the two methods is better than that of the separate ratio estimation.
【作者单位】： 农业部规划设计研究院;
【基金】：国家自然科学基金(编号:41301506) 国家重点研发计划(编号:2016YFB0501505)
【分类号】：S127;S511
，

本文编号：1817615