为了解决土壤采样中精度与经济性的平衡问题,利用计算机模拟采样研究了规则网格土壤采样时合理的采样点密度.首先构造了一个数学扩散模型,设置2～4个种子在一个100×100网格(1×1单位)的不同地方,根据扩散模型进行扩散和叠加,生成模拟的土壤属性分布地图,其结果可很好地模拟某些土壤属性的分布.利用计算机按照不同的网格单元尺寸(如3×3,5×5,7×7等)进行采样,之后利用采样值进行IDW插值处理,将数据点恢复到原始的10000个,并把插值结果与原始值进行比较即可得到采样误差.研究结果表明,当采样网格单元尺寸为属性地图输出栅格单元尺寸的11倍和17倍时,相对采样误差分别为10%和15%.合理的采样密度可以根据允许的采样误差及要求的属性地图输出栅格单元尺寸而定.%The aligned grid sampling pattern is one of the most convenient sampling methods in precision agriculture. The objective of this research is to use computer simulation to determine the optimal sampling density for the aligned grid soil sampling pattern to balance the cost and precision requirements in precision agriculture applications. A mathematical diffusion model was constructed, and 2～4 diffusion seeds were set in a 100×100 grid with a grid cell size of 1×1 unit to diffuse and accumulate point values according to the diffusion model. The diffusion results could well simulate the distribution of some attributes in the field. The computer simulated sampling points with different grid cell sizes were generated and then interpolated into 10000 points with inverse distance weighted(IDW). Finally, the original value was compared with the interpolated value for the same grid cell to find the sampling error, and optimal sampling points were determined according to the allowable sampling error. Test results showed that if the sampling grid cell sizes were about 11 times and 17 times the output cell size of the attribute map, the average relative sampling errors could be limited to 10% and 15%, respectively. The optimal density of sampling points can be determined by the sampling grid cell size and the field area.
展开▼
