With consideration of highly-efficient use of natural resources, reducing production cost and construction of high-standard agricultural fields, rice production of good seed＋fertilizer investment is reformed on basis...With consideration of highly-efficient use of natural resources, reducing production cost and construction of high-standard agricultural fields, rice production of good seed＋fertilizer investment is reformed on basis of research of smash-ridging technology, and rice smash-ridging ecological and highly-efficient cultivation was pro- posed, namely, smash-ridging based deeply poughing and rotary tillage technology was used to loosen soils deeply, with the depth from 13-15 cm to 26-28 cm. Fur- thermore, after soils softening, seedling slinging or direct seeding was adopted, which is dominated by natural rainfall and supplemented by artificial irrigation. The test proved that the technology help increasing yield and profits over 20%, with chemical fertilizer reduced by more than 10%, and labor cost reduced by 10%. What＇s more, if the technology applied once, no-tillage or slight tillage is recom- mended in the field, which would maintain original eco-conditions of soils and reach high yield, with energy, water, labor costs reduced in an environment-friendly way.展开更多
【目的】利用土壤近表面空气温湿度与土壤内部参数的关联关系对耕作层土壤水分、温度进行精准预测,为实现精细化农业种植管理提供服务。【方法】针对土壤耕作层水分、温度预测在训练集获取与模型验证等方面的实际需求,设计了基于嵌入式...【目的】利用土壤近表面空气温湿度与土壤内部参数的关联关系对耕作层土壤水分、温度进行精准预测,为实现精细化农业种植管理提供服务。【方法】针对土壤耕作层水分、温度预测在训练集获取与模型验证等方面的实际需求,设计了基于嵌入式系统及窄带物联网(Narrow band internet of things,NB-IoT)无线通信技术的物联网数据采集系统。在此基础上基于深度Q学习(Deep Q network,DQN)算法探索了一种模型组合策略,以长短期记忆(Long short-term memory,LSTM)、门限循环单元(Gated recurrent unit,GRU)与双向长短期记忆网络(Bidirectional long short-term memory,Bi-LSTM)为基础模型进行加权组合,获得了DQN-L-G-B组合预测模型。【结果】数据采集系统实现了对等间隔时间序列环境数据的长时间稳定可靠采集,可以为基于深度学习的土壤水分、温度时间序列预测工作提供准确的训练集与验证集数据。相对于LSTM、Bi-LSTM、GRU、L-G-B等模型,DQN-L-G-B组合模型在2种土壤类型(壤土、砂土)耕作层上水分与温度预测中的均方根误差(Root mean square error,RMSE)、平均绝对误差(Mean absolute error,MAE)、平均百分比误差(Mean absolute percentage error,MAPE)都有一定程度的降低,R2提高了约0.1%。【结论】通过该物联网数据采集系统与DNQ-L-G-B组合模型,可以有效地完成基于土壤近表面空气温、湿度对耕作层土壤中水分、温度的精准预测。展开更多
基金Supported by Fundamental Research Funds for Guangxi Academy of Agricultural Sciences(2014YZ07)Transformation Project of Scientific and Technological Achievements,Guangxi Academy of Agricultural Sciences(201405)~~
文摘With consideration of highly-efficient use of natural resources, reducing production cost and construction of high-standard agricultural fields, rice production of good seed＋fertilizer investment is reformed on basis of research of smash-ridging technology, and rice smash-ridging ecological and highly-efficient cultivation was pro- posed, namely, smash-ridging based deeply poughing and rotary tillage technology was used to loosen soils deeply, with the depth from 13-15 cm to 26-28 cm. Fur- thermore, after soils softening, seedling slinging or direct seeding was adopted, which is dominated by natural rainfall and supplemented by artificial irrigation. The test proved that the technology help increasing yield and profits over 20%, with chemical fertilizer reduced by more than 10%, and labor cost reduced by 10%. What＇s more, if the technology applied once, no-tillage or slight tillage is recom- mended in the field, which would maintain original eco-conditions of soils and reach high yield, with energy, water, labor costs reduced in an environment-friendly way.
文摘【目的】利用土壤近表面空气温湿度与土壤内部参数的关联关系对耕作层土壤水分、温度进行精准预测,为实现精细化农业种植管理提供服务。【方法】针对土壤耕作层水分、温度预测在训练集获取与模型验证等方面的实际需求,设计了基于嵌入式系统及窄带物联网(Narrow band internet of things,NB-IoT)无线通信技术的物联网数据采集系统。在此基础上基于深度Q学习(Deep Q network,DQN)算法探索了一种模型组合策略,以长短期记忆(Long short-term memory,LSTM)、门限循环单元(Gated recurrent unit,GRU)与双向长短期记忆网络(Bidirectional long short-term memory,Bi-LSTM)为基础模型进行加权组合,获得了DQN-L-G-B组合预测模型。【结果】数据采集系统实现了对等间隔时间序列环境数据的长时间稳定可靠采集,可以为基于深度学习的土壤水分、温度时间序列预测工作提供准确的训练集与验证集数据。相对于LSTM、Bi-LSTM、GRU、L-G-B等模型,DQN-L-G-B组合模型在2种土壤类型(壤土、砂土)耕作层上水分与温度预测中的均方根误差(Root mean square error,RMSE)、平均绝对误差(Mean absolute error,MAE)、平均百分比误差(Mean absolute percentage error,MAPE)都有一定程度的降低,R2提高了约0.1%。【结论】通过该物联网数据采集系统与DNQ-L-G-B组合模型,可以有效地完成基于土壤近表面空气温、湿度对耕作层土壤中水分、温度的精准预测。
