基于人工神经网络的土壤养分肥力等级评价方法

Soil Fertility Evaluation Based on BP Artificial Neural Network

土壤养分肥力等级是土壤特性的综合反映,也是揭示土壤条件动态的最敏感的指标。对它的评价涉及多个指标,很难用常规方法进行。人工神经网络由于有大规模并行处理、分布式存储、自适应性、容错性等特点,可以解决复杂的非线性高维同题,本研究拟采用建立的土壤养分肥力等级BP神经网络评价模型,对山西省大同地区的土壤养分肥力等级进行评价。

Soil fertility includes three aspects, soil fertilizer quality, soil environment quality and soil healthy quality. It is a reflection of soil characteristics, but it is difficult to predict with common methods because it is determined by several parameters. Artificial neural networks are one of advanced courses having many merits, thus it was applied to soil quality and quantity prediction. Artificial neural network and BP neural networks were interpreted, and 7-5-1 of BP neural networks model was established to study soil fertility degree of DATONG, Shanxi Province. The results showed that artificial neural networks could be applied well to soil fertility evaluation and prediction.