基于物联网系统的甘蔗田间环境数据聚类分析与评价

Cluster Analysis and Evaluation of Sugarcane Field Environmental Data Based on IoT System

针对甘蔗田间环境数据进行数据挖掘,从而为甘蔗的生长提供科学的指导意见,并为后期甘蔗产量与品质的预测提供技术支持。本研究采用了NXP公司的JN5168无线模块构建了一套田间物联网,可采集田间的空气温湿度、光照强度、土壤湿度、土壤温度等数据。并对一甘蔗试验地采集到的田间环境数据进行了分析与评价。首先采用差值法对环境数据进行了预处理以剔除异常数据,然后运用层次聚类法对甘蔗不同生长期的环境数据进行聚类,并根据甘蔗生长环境的适宜条件对聚类结果进行了好中差三个等级的评价。实验结果表明,甘蔗萌芽期、幼苗期、分蘖期和伸长期的层次聚类结果的同表象相关系数分别为0.849 1、0.835 5、0.823 9和0.817 5,说明层次聚类在分类方面效果较好。

This paper studied on data mining of sugarcane field environmental data,so to provide scientific guidance for sugarcane growth as well as technical support for the prediction of sugarcane yield and quality in the later period.In this study,JN5168 wireless module from NXP Corporate was used to design a field Io T system to collect environmental data such as air temperature and humidity,illumination,soil moisture and soil temperature. The environmental data collected in a sugarcane experimental field were then analyzed and evaluated.Difference method was first used to preprocess the environmental data to remove abnormal data,then the hierarchical clustering method was used to cluster the environmental data of different growth stages,meanwhile the clustering results were evaluated into three categories in accordance with the suitable growth environment of sugarcane. The cophenetic correlation coefficient of the hierarchical clustering results in germination stage,seedling stage,tillering stage and elongation stage were 0.849 1,0.835 5,0.823 9 and 0.817 5,respectively,indicating that hierarchical clustering worked well in data classification.