基于LSTNet的果园土壤含水量预测模型研究
Research on prediction model of soil water content in orchard based on LSTNet
彭东 1周建平 2许燕 2彭炫 2吴昊臻 1秦春雨1
作者信息
- 1. 新疆大学机械工程学院,新疆乌鲁木齐830000
- 2. 新疆大学机械工程学院,新疆乌鲁木齐830000;新疆维吾尔自治区农牧机器人及智能装备工程研究中心,新疆乌鲁木齐830000
- 折叠
摘要
为了有效利用核桃园各传感器数据,预测核桃园土壤含水量来指导灌溉,建立了基于LSTNet的果园土壤含水量预测模型.首先,确定LSTNet预测模型输入参数;其次,对LSTNet预测模型进行多特征和单一特征输入对比分析;最后,将LSTNet预测模与LSTM预测模型和卷积神经网络(CNN)预测模型进行预测精度比较分析.研究结果表明:LSTNet预测模型相较于其他二种预测模型在果园土壤含水量预测上有更高的预测精度.
Abstract
In order to effectively utilize the sensor data of walnut orchard to predict the soil water content of walnut orchard to guide irrigation,a prediction model of orchard soil water content based on LSTNet is established.Firstly,the input parameters of LSTNet prediction model are determined.Secondly,the multi-feature and single feature input of LSTNet prediction model are compared and analyzed.Finally,the prediction precision of LSTNet prediction model is compared and analyzed with LSTM prediction model and CNN prediction model.The research results show that compared with the other two prediction models,LSTNet prediction model has higher prediction precision in predicting soil water content in orchard.
关键词
核桃园/土壤/预测/神经网络/含水量Key words
walnut orchard/soil/prediction/neural network/water content引用本文复制引用
基金项目
新疆维吾尔族自治区创新团队项目(2022D14002)
出版年
2024
NETL
NSTL
万方数据