基于3S 的森林植被面积空间抽样方法

摘要

以辽宁省为例，应用中分辨率成像光谱仪影像（ MODIS）数据进行森林信息提取，并选取森林面积为研究对象，采用网格单元作为抽样单元，应用随机、系统、分层的抽样方法对辽宁省森林面积进行抽样估算，在获得最优抽样方法的基础上进行第2阶段抽样，设计不同样本容量水平下的森林面积空间抽样方法试验。结果表明：以县森林面积覆盖百分比为分层标志的分层抽样方法抽样效率最高，在第2阶段抽样估算的8种样本容量水平下，当样本容量为8时，抽样精度达到95％，并且所需样本数量最少，达到了森林面积抽样精度要求。应用此方法能够减少森林资源调查成本和工作量，可以结合遥感、地理信息系统和全球定位系统技术（3S技术），进行大区域范围森林面积的监测。%Taking Liaoning Province as a case and forest area in the province as the research object , which was extracted by applying MODIS data, adopting grid cells as the sampling unit , the methods of random sampling , systematic sampling and stratified sampling were conducted for sampling assumption on the forest area of the whole province .With the optimal sampling method , the second-stage sampling was facilitated , and experiments on forest area spatial sampling method was designed in different sample sizes.The sampling efficiency of stratified sampling with the percentage of forest cover of the counties as stratified symbol was the highest.Under the 8 sample sizes sampled and assumed in the second-stage sampling, when the sample size was 8 the sampling accuracy reached 95%and it needed the fewest samples .It satisfied the requirements of forest area sampling accuracy .This method reduced cost and workload of the forest resource investigation and provided reference for large regional forest area monitoring with remote sensing, geographic information system and global positioning system (3S technology).