基于大数据统计的土壤养分含量预测模型设计

Design of soil nutrient content prediction model based on big data statistics

针对传统预测模型预测土壤养分含量精度低的问题,设计基于大数据统计的土壤养分含量预测模型。使用光谱采集仪采集土壤养分含量数据,对数据进行平滑、标准化和正交化处理,消除影响预测精的数据。对处理后的光谱图进行分析,总结土壤养分大数据规律,明确在反射曲线突变处的不同波段对应不同含量的土壤养分。重复多次后,完成对模型精度的校验,实现基于大数据统计的土壤养分含量预测模型的设计。通过与传统模型的对比实验,验证设计的模型能够提高大约3倍的预测精度,更适用于对土壤养分含量进行预测。

A soil nutrient content prediction model based on big data statistics is designed to improve the low accuracy of traditional prediction model in predicting soil nutrient content. The data of soil nutrient content is collected with the spectral acquisition instrument,and the data is smoothed,standardized and normalized to eliminate the influence of prediction precision.The processed spectrogram is analyzed to summarize the rule of big data of soil nutrients,and it is clear that different bands at the abrupt change of reflection curve correspond to different contents of soil nutrients. The checking of model precision is completed after repeated for many times,and the design of soil nutrient content prediction model based on big data statistics is realized. The experimental results verify that,in comparison with the traditional model,the designed model can increase the prediction accuracy by about 3 times,and is more suitable for the prediction of soil nutrient content.