Design of the Spatial Sampling Scheme for Estimating the Cultivation Area of Garlic and Onion Using Satellite-Based and Unmanned Aerial Vehicle Remotely Sensed Data

Abstract

Garlic and onion are the most important vegetable crops in Korea. The accurate estimation of cultivation area for garlic and onion acreage is critical to predict the production of vegetable crops, adjust agricultural planting plan and ensure an effective supply of farm products. However, the plots cultivated with garlic and onion are very fragmentary and dispersive because the cross cropping commonly occurs in the two crops. Therefore, it is very difficult to establish the accurate identification of garlic and onion using satellite-based remotely sensed images alone. In case of tracking Hapcheon Gun, Korea as the sampling site, objective of this study was to formulate a spatial sampling scheme through combining satellite-based, unmanned aerial vehicle (UAV) remotely sensed images and the stratified sampling method for improving the estimated accuracy to cultivation area of garlic and onion. The results are shown that there was almost no classification error, when UAV remotely sensed image was used to retrieve the cultivation area of garlic and onion. The error found in the two crops classification using Rapid Eye satellite-based images, and the classification error for garlic was the larger than that of onion;. The variance for cultivation area of garlic and onion within each stratum can be significantly decreased, when the proportion of the cultivation area for two crops is accounted for one sampling unit. It was observed that the required sample size for meeting the designed extrapolation accuracy decreased with the stratification number of the sampled population. Comprehensively considering population extrapolating accuracy, sampling survey cost and rationality, 10 strata was the optimum stratification number. It was appeared that the spatial stratified sampling scheme combining satellite-based and UAV remotely sensed images had a high accuracy and stability for estimation of cultivation areas for the two crops, because both the relative error and CV of population extrapolation using this scheme was less than 10%. Results of population extrapolation and error estimation for garlic acreage in the study area in 2017. Stratum number Nh N Wh nh fh (%) (m2) sh2 1 198 422 0.4692 28 0.1414 10272.57 111059383.60 2 78 422 0.1848 10 0.1282 23347.15 441781199.13 3 47 422 0.1114 6 0.1277 53480.35 217588913.35 4 41 422 0.0972 5 0.1220 72042.97 598687783.10 5 24 422 0.0569 3 0.1250 101778.39 85000902.09 6 17 422 0.0403 3 0.1765 116648.67 1006782408.64 7 7 422 0.0166 2 0.2857 143419.00 765673525.59 8 4 422 0.0095 2 0.5000 144819.77 2796755176.20 9 3 422 0.0071 2 0.6667 269129.69 10129222962.02 10 3 422 0.0071 2 0.6667 261057.10 1065556621.53 r (%) 7.22 CV (%) 5.12