基于CNN的高校节水智能灌溉设计

Design of water-saving irrigation in colleges based on CNN

针对高校后勤建设中水资源利用率较低、灌溉管理过于粗放等问题,文章提出了一种基于CNN的智能灌溉设计思路。首先,基于ZigBee技术和5G技术,实现校园绿色植被数据信息的采集与传输、灌溉控制命令的派发以及控制终端设备的控制。其次,构建基于卷积神经网络和自回归模型的温湿度数据预测模型。该模型选择相关度较高的有效降水量、作物需水量、最高气温、最低气温4个因子进行训练。实验结果表明,模型的预测结果很好的拟合真实值变化,得到的作物灌溉用水量与真实灌溉量相差不大,可满足作物需水量的预测要求。

Aimed at the problems of low water resource utilization and extensive irrigation management in the logistics construction of colleges.This paper proposes an intelligent irrigation design idea based on CNN.Firstly,based on ZigBee technology and 5G technology,the collection and transmission of campus green vegetation data information,the distribution of irrigation control commands and the control of terminal equipment are realized.Secondly,a temperature and humidity data prediction model based on convolutional neural network and autoregressive model is constructed.Four factors of effective precipitation,crop water requirement,maximum temperature and minimum temperature with high correlation were selected to train the model.The experimental results show that the prediction results of the model fit the real value changes well,and the obtained irrigation water consumption is not different from the real irrigation amount,which meets the prediction requirements of crop water demand.