Prediction models of soil moisture content and electrical conductivity in citrus orchard based on internet of things and LSTM
-
摘要:目的
构建柑橘果园环境信息物联网实时采集系统,建立基于物联网和长短期记忆(LSTM)的柑橘园土壤含量和电导率预测模型,为果园灌溉施肥管理、效果预测评估提供参考依据。
方法利用土壤温度、含水量、电导率三合一传感器,在柑橘果园中设置5个节点和1个气象站,通过ZigBee短距离无线通信和GPRS远距离无线传输,将果园气象数据和土壤墒情数据传输至远程服务器。利用LSTM模型建立气象数据与土壤含水量和电导率的预测模型,计算均方根误差(RMSE)和决定系数(R2)以进行性能评估。
结果物联网系统能够实现远程传输柑橘果园环境数据,建立了基于LSTM和广义回归神经网络(GRNN)的土壤含水量和电导率预测模型,模型在5个节点的数据集的训练结果分别为:LSTM模型训练的土壤含水量和电导率的RMSE范围分别为6.74~8.65和6.68~8.50,GRNN模型训练的土壤含水量和电导率的RMSE范围分别为7.01~14.70和7.60~13.70。利用生成的LSTM模型和气象数据进行拟合,将土壤含水量和电导率的预测值与实测值进行回归分析,LSTM模型拟合的土壤含水量和电导率的R2范围分别为0.760~0.906和0.648~0.850,GRNN模型拟合的土壤含水量和电导率的R2范围分别为0.126~0.369和0.132~0.268,说明LSTM模型的性能表现较好。
结论建立了柑橘果园环境的物联网信息传输系统,构建的基于LSTM的果园土壤含水量和电导率预测模型具有较高的精度,可用于指导柑橘果园的灌溉施肥管理。
Abstract:ObjectiveTo build an internet of things (IoT) system for transmitting the environmental information of citrus orchards in real time, establish prediction models of soil moisture content and electrical conductivity in citrus orchard based on IoT system and long short-term memory (LSTM), and provide references for irrigation and fertilization management as well as effect prediction and evaluation.
MethodSoil temperature, moisture and electrical conductivity sensors were applied in five IoT nodes and a weather station was set in citrus orchard. The meteorological data and soil moisture data collected in the orchard were transmitted to a remote server via ZigBee, a short range wireless communication technique, and GPRS, a long distance wireless transmission technique. The prediction models of soil moisture content and electrical conductivity were established using weather data based on the LSTM model. The root mean square error (RMSE) and coefficient of determination (R2) were calculated to evaluate the performance of the model.
ResultThe IoT system was capable to transmit environmental data of the citrus orchard to a remote server. LSTM and general regression neural network (GRNN) model were built to predict soil moisture content and electrical conductivity. The performance of models in five nodes were as following: The RMSE of soil moisture content and electrical conductivity ranged from 6.74 to 8.65 and 6.68 to 8.50 respectively based on LSTM model, and ranged from 7.01 to 14.70 and 7.60 to 13.70 respectively based on GRNN model. With the generated LSTM model and meteorological data for predicting, regression analysis was conducted between predicted and measured values of soil moisture content and electrical conductivity. The R2 of soil moisture content and electrical conductivity ranged from 0.760 to 0.906 and 0.648 to 0.850 respectively based on LSTM model, and ranged from 0.126 to 0.369 and 0.132 to 0.268 respectively based on GRNN model. The results indicated that the LSTM model performed better than the GRNN model.
ConclusionThe IoT system for citrus orchard environmental information transmission is established. The LSTM model has high accuracy in predicting soil moisture content and electrical conductivity, and the model can be helpful for guiding irrigation and fertilization management.
-
-
表 1 土壤含水量和电导率与环境参数的相关系数
Table 1 Correlation coefficients of soil moisture, electrical conductivity and environmental parameters
土壤因子
Soil factor空气温度 Air temperature 降雨量
Precipitation空气湿度
Air
humidity土壤温度
Soil
temperature蒸散量
Evapotranspiration最高值
Max.
最低值
Min.
平均值
Mean
含水量 Moisture content 0.490 0.598 0.644 0.123 0.430 0.652 0.158 电导率 Electrical conductivity 0.467 0.311 0.231 0.300 0.168 0.251 0.148 表 2 节点数据信息
Table 2 Datasets of nodes
节点
Node数据总数
Total number训练集
Training dataset测试集
Testing dataset1 216 151 65 2 241 168 73 3 241 168 73 4 196 137 59 5 203 142 61 气象站
Weather station241 168 73 表 3 模型性能表现
Table 3 Performance of models
节点
Node模型
Model土壤含水量
Soil moisture
content土壤电导率
Soil electrical
conductivityRMSE R2 RMSE R2 1 LSTM 8.65 0.811 6.88 0.850 GRNN 13.30 0.126 12.30 0.106 2 LSTM 7.30 0.906 6.68 0.727 GRNN 14.70 0.323 13.70 0.185 3 LSTM 6.74 0.814 8.22 0.648 GRNN 7.01 0.369 8.26 0.268 4 LSTM 7.93 0.760 8.50 0.767 GRNN 10.27 0.256 7.85 0.167 5 LSTM 8.40 0.768 7.78 0.769 GRNN 8.80 0.206 7.60 0.132 -
[1] 成家壮, 韦小燕, 范怀忠. 广东柑橘疫霉研究[J]. 华南农业大学学报, 2004, 25(2): 31-33. [2] 张超博, 李有芳, 李思静, 等. 土壤管理方式对伏旱期柑橘生长及土壤温度和水分的影响[J]. 华南农业大学学报, 2019, 40(3): 45-52. [3] 肖畅, 彭婷, 刘继红. 基于WOS和CiteSpace分析我国近十年柑橘研究热点与前沿[J]. 果树学报, 2020.doi: 10.13925/j.cnki.gsxb.20200119. [4] 李莉萍. 柑橘产业化经营案例分析: 评《柑橘绿色生产与产业化经营》[J]. 中国果树, 2020(3): 151. [5] 罗省根, 双巧云, 陈团显, 等. 江西省柑橘出口现状及发展对策[J]. 中国果树, 2020(3): 138-140. [6] 赖俊桂, 孙道宗, 王卫星, 等. 基于无线传感器网络的山地柑橘园灌溉控制系统设计与试验[J]. 江苏农业科学, 2020, 48(7): 245-249. [7] 陈昱辛, 贾悦, 崔宁博, 等. 滴灌水肥一体化对柑橘叶片光合、产量及水分利用效率的影响[J]. 灌溉排水学报, 2018, 37(S2): 50-58. [8] 谢家兴, 高鹏, 莫昊凡, 等. 荔枝园智能灌溉决策系统模糊控制器设计与优化[J]. 农业机械学报, 2018, 49(8): 26-32. [9] 赖呈纯, 黄贤贵, 王琦, 等. 果实生长与果园土壤含水量的变化对‘茂谷柑’裂果的影响[J]. 福建农林大学学报(自然科学版), 2019, 48(4): 434-439. [10] 朱浩, 刘珂欣, 刘维维, 等. 极端耐盐碱菌株的筛选及其菌肥对盐碱条件下小麦生长和土壤环境的影响[J]. 应用生态学报, 2019, 30(7): 2338-2344. [11] 杨伟志, 孙道宗, 刘建梅, 等. 基于物联网和人工智能的柑橘灌溉专家系统[J]. 节水灌溉, 2019(9): 116-120. [12] 李中良, 胡晨晓, 邹腾飞, 等. 基于物联网的柑橘土壤水分养分实时监测系统的设计与实现[J]. 农业网络信息, 2014(2): 21-24. [13] 罗党, 王浍婷. 灰色神经网络下的多变量土壤含水量预测模型[J]. 华北水利水电大学学报(自然科学版), 2017, 38(5): 70-75. [14] 蔡亮红, 丁建丽. 基于变量优选和ELM算法的土壤含水量预测研究[J]. 光谱学与光谱分析, 2018, 38(7): 2209-2214. [15] 张武, 洪汛, 李蒙, 等. 监测采样间隔对土壤墒情预测模型性能的影响[J]. 甘肃农业大学学报, 2020, 55(1): 221-228. [16] 何灿隆, 沈明霞, 刘龙申, 等. 基于NB-IoT的温室温度智能调控系统设计与实现[J]. 华南农业大学学报, 2018, 39(2): 117-124. [17] 岳学军, 王叶夫, 刘永鑫, 等. 基于GPRS与ZigBee的果园环境监测系统[J]. 华南农业大学学报, 2014, 35(4): 109-113. [18] 徐兴, 岳学军, 林涛. 基于ZigBee网络的水环境无线监测系统设计[J]. 华南农业大学学报, 2013, 34(4): 593-597. [19] 余国雄, 王卫星, 谢家兴, 等. 基于物联网的荔枝园信息获取与智能灌溉专家决策系统[J]. 农业工程学报, 2016, 32(20): 144-152. [20] 姜晟, 王卫星, 孙道宗, 等. 能量自给的果园信息采集无线传感器网络节点设计[J]. 农业工程学报, 2012, 28(9): 153-158. [21] 王卫星, 陈华强, 姜晟, 等. 基于低功耗的发射功率自适应水稻田WSN监测系统[J]. 农业机械学报, 2018, 49(3): 150-157. [22] ALLEN R G, PEREIRA L S, RAES D, et al. Crop evapotranspiration: Guidelines for computing crop water requirements: FAO irrigation and drainage paper 56[EB/OL]. [2020-06-30]. https://www.researchgate.net/publication/235704197.
[23] RALLO G, GONZÁLEZ-ALTOZANO P, MANZANO-JUÁREZ J, et al. Using field measurements and FAO-56 model to assess the eco-physiological response of citrus orchards under regulated deficit irrigation[J]. Agr Water Manag, 2017, 180: 136-147. doi: 10.1016/j.agwat.2016.11.011
[24] FENG Y, PENG Y, CUI N, et al. Modeling reference evapotranspiration using extreme learning machine and generalized regression neural network only with temperature data[J]. Comput Electron Agr, 2017, 136: 71-78. doi: 10.1016/j.compag.2017.01.027
[25] SPECHT D F. A general regression neural network[J]. IEEE Trans Neural Netw, 1991, 2(6): 568-576. doi: 10.1109/72.97934
[26] MAJHI B, NAIDU D, MISHRA A P, et al. Improved prediction of daily pan evaporation using Deep-LSTM model[J]. Neural Comput Appl, 2019, 4(6): 110-122.
[27] ZHANG J, ZHU Y, ZHANG X, et al. Developing a long short-term memory (LSTM) based model for predicting water table depth in agricultural areas[J]. J Hydrol, 2018, 561: 918-929. doi: 10.1016/j.jhydrol.2018.04.065
[28] CHENG H, XIE Z, WU L, et al. Data prediction model in wireless sensor networks based on bidirectional LSTM[J]. Eurasip J Wirel Comm, 2019, 2019(1): 203-212. doi: 10.1186/s13638-019-1511-4
[29] MAJHI B, NAIDU D, MISHRA A P, et al. Improved prediction of daily pan evaporation using Deep-LSTM model[J]. Neural Comput Appl, 2020, 32(12): 7823-7838.
[30] REDDY D S, PRASAD P R C. Prediction of vegetation dynamics using NDVI time series data and LSTM[J]. Model Earth Syst, 2018, 4(1): 409-419. doi: 10.1007/s40808-018-0431-3
[31] NASH J E, SUTCLIFFE J V. River flow forecasting through conceptual models part I: A discussion of principles[J]. J Hydrol, 1970, 10(3): 282-290. doi: 10.1016/0022-1694(70)90255-6