论文部分内容阅读
将吉林省榆树市弓棚镇作为研究区域,利用采样密度和插值方法(BP神经网络方法和Kriging方法),研究其对农田土壤碱解氮空间变异性的影响。研究结果表明:Kriging模型插值精度随采样密度的减少,呈显著下降趋势,采样密度对BP神经网络插值精度影响相对较小。当采样密度较大时,插值精度表现为Kriging模型显著高于BP神经网络模型;当采样密度较小时,BP神经网络模型的插值精度优于Kriging模型插值精度。研究为精准农业土壤养分插值方法的选取、制定优化采样策略提供科学依据。
Taking Gongspeng Town of Yushu City, Jilin Province as a research area, the effects of soil available nitrogen on the spatial variability of soil available nitrogen (N 2 O 3 N) were studied using sampling density and interpolation (BP neural network and Kriging methods). The results show that the interpolation accuracy of Kriging model decreases with the decrease of sampling density, and the sampling density has a relatively small influence on the interpolation accuracy of BP neural network. When the sampling density is large, the interpolation accuracy is significantly higher than that of the BP neural network model for Kriging model. When the sampling density is small, the interpolation accuracy of the BP neural network model is better than that of the Kriging model. The research provides a scientific basis for the selection of precision agricultural soil nutrient interpolation methods and for the development of optimal sampling strategies.