基于无人机样方事后分层的作物面积估算

无人机是开展野外调查的一种新型、有效的手段,能够及时、准确地获取地面调查样方信息,为作物面积遥感估算提供精准的样方数据。研究针对无人机抽样调查的样方特点,提出了适合于无人机样方的多层次事后分层指标(多层次-异质性指标、多层次-面积规模指标)。将这些指标分别用于事后分层抽样,估算冬小麦面积。并根据变异系数(coefficient of variance,CV)对其抽样效率进行评价。多层次指标是将多种分层指标分层结果叠加形成的,能够充分反映作物种植空间分布特征、空间异质性及种植规模,可以保证作物种植面积遥感估算的精度。以河南省冬小麦面积估算为例,在冬小麦空间分布空间范围,建立300m×300m抽样框格网,作为抽样基本单元。分别利用实验设计的多层次-面积规模指标、多层次-异质性指标、面积规模指标、异质性指标计算各抽样基本单元的对应指标值。按照累计平方根法计算不同分层指标下的分层界限值。最后进行事后分层估计,计算分层效率,对分层结果与分层效率进行对比分析。计算得到以上4种分层方法的变异系数分别为1.85%、1.41%、2.16%、1.55%。结果表明,结合无人机抽样调查,利用多层次指标法进行分层,各层内作物均质性较好,能够提高农作物面积估算的精度;此外,异质性指标较面积规模指标更能提高分层的层内作物均质性与农作物面积估算精度。

STUDY ON THE POST-STRATIFICATION METHOD FOR CROP AREA ESTIMATION BASED ON UNMANNED AERIAL VEHICLE IMAGES

Unmanned aerial vehicle ( UAV ) is a new effective approach for field survey and capable of getting ground survey plots information timely and accurately, which can provide the accurate sampling data for crop area estimation. This paper proposed two multi-level indices for post-stratification ( multi-level heterogeneity index and multi-level acreage index) which were suitable for the UAV quadrat sampling. And then the indices were used for the estimation of winter wheat area by stratified sampling. The CV ( coefficient of variance) was used to as-sess the efficiency of sampling. The multi-level index was effective to improve the accuracy of estimation of acre-age and can adequately represent spatial distribution of planting and planting scale. Taking the estimation of the ar-ea of wheat in Henan province as an example, this paper built 300m × 300m grids as the primary sample unit ( PSU) covering the range of whiter wheat, and then calculated the indices value of each unit using 4 multi-level indices,i. e. , multi-level heterogeneity index, multi-level acreage index, acreage index and heterogeneity in-dex. The indices reflected the features of spatial distribution and complexity of the structure of crop, respectively. And then it calculated the value of stratified boundary using cumulative square root method. Finally, it estimated the area of winter wheat based on post-stratification and analyzed the result of efficiency of estimation. The results showed that the CVs were 1. 85%、1. 41%、2. 16%、1. 55% , respectively, which indicated that the multi-index can be an effective strata in post-stratification for the UAV's quadrat ground truth data. It can represented the crop size, spatial heterogeneity and the degree of complexity of crop planting structure, and improve the accuracy of esti-mation of acreage of wheat with the UAV sampling techniques. In addition, comparing with the acreage index,the heterogeneity index was better for improving the crop homogeneity in each strata and the estimation accuracy of crop acreage.