基于Landsat8 OLI遥感影像和森林资源二类调查数据,选择云南省普洱市镇沅县为研究区,在90%的抽样精度(E)与95%的可靠性指标(t_α)的条件下,应用二阶抽样技术,以镇沅县思茅松单位面积AGB、单位面积地上蓄积量、7个单波段及5个植被指数作为抽样标志,对不同抽样方案的抽样总体方差、变动系数、标准误差、绝对误差、估测精度、AGB估测值及估测误差进行分析,并与简单随机抽样、系统抽样对比,分析不同抽样方法应用不同抽样标志的综合效率。结果表明:1)二阶抽样综合效率远远高于简单抽样和系统抽样;2)基于单波段和植被指数的二阶抽样效率普遍优于基于二调数据的,二阶抽样效率最好的抽样标志是ARVI与NDVI,2种植被指数仅需154个样本,较基于二调数据的二阶抽样降低60%的样本量,精度能达到最高,分别为97.50%和97.67%。研究结果说明基于遥感因子的二阶抽样可以大幅提高抽样效率。

Based on Landsat8 OLI remote sensing imagery and the second class survey data of forest resources,taking Zhenyuan County,Pu'er City,Yunnan Province as the study area,with sampling accuracy (E) of 90% and reliability index (t_α) of 95%,the two-stage sampling technique was applied,using AGB per unit area,above-ground accumulation per unit area,seven single bands and five vegetation indices of Simao pine in Zhenyuan County as the sampling markers,so that the overall sampling variance,coefficient of variation,standard error,absolute error,estimation accuracy,AGB estimation value and estimation error of different sampling schemes were analyzed and compared with simple random sampling and systematic sampling to analyze the comprehensive efficiency of different sampling methods applying different sampling marks.The results showed that:1) the comprehensive efficiency of the two-stage sampling was much higher than that of simple sampling and systematic sampling,2) The efficiency of the two-stage sampling based on single band and vegetation indices was generally better than which was based on the 2nd-class survey data,and the best sampling signs of the two-stage sampling efficiency were ARVI and NDVI,and only 154 samples were required for the two vegetation indices,which reduced the sample size by 60% compared with the two-stage sampling based on the 2nd-class survey data,and the precision could reach the highest,which were 97.50% and 97.67%,respectively.The two-stage sampling based on remote sensing factors could significantly improve the sampling efficiency.