基于WDNN的温室多特征数据融合方法研究
|
摘要
目前物联网监测产品在温室生产中大量应用产生海量数据,但现有用于温室数据融合算法对高维特征及混合特征(数据同时包含稀疏特征和连续特征)处理精度较低且泛化能力较弱,鲜有利用深度学习模型对温室数据进行顶层融合并提供准确的决策信息。本文提出了一种基于宽-深神经网络(Wide-deep neural network,WDNN)的两级温室环境数据融合算法。首先利用温室内多点多特征数据训练WDNN深度学习模型,输出形式为多点单特征,再将该输出数据按照少数服从多数原则进行融合,得到温室环境状态的整体评估结果。试验结果表明,该融合方法对预测集中混合特征的决策准确率达到98. 90%,融合特征类型的增加,可用于监测参数更多、环境更复杂的温室,将WDNN模型用于温室混合数据融合是可行有效的,在保证决策精度的同时丰富了可融合特征类别,进一步提升温室融合系统的智能化程度,对温室智能调控提供有效技术支撑。
The IoT monitoring products are widely used in greenhouse production,which could generate massive data. The existing data fusion algorithms for greenhouses had low fusion accuracy and weak generalization capability for high-dimensional features and mixed features( combined with sparse features and continuous features). It was rare to use the deep learning model to top-level fusion of greenhouse data and provide accurate decision information. Aiming at the above problems,a two-level greenhouse environment data fusion algorithm was proposed based on wide-deep neural network( WDNN). Firstly,integrating multi-point multi-features mixed data in the greenhouse and marking the data categories. Then the constructed training set and test set were input into the WDNN deep learning model for 2000-step iteration training. The model structure was set as 7-100-50-7,the output form was multi-point single feature,which was the first-level fusion result as decision information of each area of the greenhouse,and then the output data was second-level fusion according to the minority obeyed majority principle,and the overall evaluation decision of the greenhouse environmental state was obtained. For comparison purposes,the other three fusion models were trained as deep neutral network( DNN),BP neural network( BPNN)and random forest( RF). The experimental results showed that the loss value of the initial segment of the WDNN network was higher than that of DNN network,but the loss function curve had a faster rate of decline and the model parameters were better. The accuracy of the model after training was 4. 32 percentage points higher than that of DNN,but the training time was increased by 21. 36%; the accuracy of BPNN model was 82% and its parameter optimization was the slowest,parameter optimization required more iteration steps; RF model training speed was the fastest,but its model fusion accuracy was 3. 39 percentage points lower than that of WDNN. The fusion accuracy was insufficient; above comparison results proved that it was feasible and excellent to use the WDNN model to fuse the mixed data in the greenhouse. Inputting the mixed situation information contained the sensor anomaly and meteorological data under various conditions into the fusion system,then the context decision rate reached 98. 90%. The realization of the WDNN fusion system could be used to monitor greenhouses with more parameters and more complex environments,and enrich the fusion feature categories while ensuring the accuracy of decision-making. It could further improve the intelligence degree of the greenhouse fusion system.
The IoT monitoring products are widely used in greenhouse production,which could generate massive data. The existing data fusion algorithms for greenhouses had low fusion accuracy and weak generalization capability for high-dimensional features and mixed features( combined with sparse features and continuous features). It was rare to use the deep learning model to top-level fusion of greenhouse data and provide accurate decision information. Aiming at the above problems,a two-level greenhouse environment data fusion algorithm was proposed based on wide-deep neural network( WDNN). Firstly,integrating multi-point multi-features mixed data in the greenhouse and marking the data categories. Then the constructed training set and test set were input into the WDNN deep learning model for 2000-step iteration training. The model structure was set as 7-100-50-7,the output form was multi-point single feature,which was the first-level fusion result as decision information of each area of the greenhouse,and then the output data was second-level fusion according to the minority obeyed majority principle,and the overall evaluation decision of the greenhouse environmental state was obtained. For comparison purposes,the other three fusion models were trained as deep neutral network( DNN),BP neural network( BPNN)and random forest( RF). The experimental results showed that the loss value of the initial segment of the WDNN network was higher than that of DNN network,but the loss function curve had a faster rate of decline and the model parameters were better. The accuracy of the model after training was 4. 32 percentage points higher than that of DNN,but the training time was increased by 21. 36%; the accuracy of BPNN model was 82% and its parameter optimization was the slowest,parameter optimization required more iteration steps; RF model training speed was the fastest,but its model fusion accuracy was 3. 39 percentage points lower than that of WDNN. The fusion accuracy was insufficient; above comparison results proved that it was feasible and excellent to use the WDNN model to fuse the mixed data in the greenhouse. Inputting the mixed situation information contained the sensor anomaly and meteorological data under various conditions into the fusion system,then the context decision rate reached 98. 90%. The realization of the WDNN fusion system could be used to monitor greenhouses with more parameters and more complex environments,and enrich the fusion feature categories while ensuring the accuracy of decision-making. It could further improve the intelligence degree of the greenhouse fusion system.
引文
[1]李志刚,刘丹丹,张小栓.基于分簇数据融合的农产品冷链温度监控方法[J/OL].农业机械学报,2017,48(8):302-308.LI Zhigang,LIU Dandan,ZHANG Xiaoshuan. Cold chain temperature monitoring method of agricultural products based onclustered data fusion[J/OL]. Transactions of the Chinese Society for Agricultural Machinery,2017,48(8):302-308. http:∥www. j-csam. org/jcsam/ch/reader/view_abstract. aspx? flag=1&file_no=20170835&journal_id=jcsam. DOI:10. 6041/j.issn. 1000-1298. 2017. 08. 035.(in Chinese)
[2]崔琳.北方日光温室无线传感器多数据融合技术的研究[D].沈阳:沈阳农业大学,2016.CUI Lin. Study on data analysis platform of northern sunlight greenhouse's monitor and control system[D]. Shenyang:ShenyangAgricultural University,2016.(in Chinese)
[3]熊迎军,沈明霞,陆明洲,等.温室无线传感器网络系统实时数据融合算法[J].农业工程学报,2012,28(23):160-166.XIONG Yingjun,SHEN Mingxia,LU Mingzhou,et al. Algorithm of real time data fusion for greenhouse WSN system[J].Transactions of the CSAE,2012,28(23):160-166.(in Chinese)
[4] SALAZAR R,LPEZ I,ROJANO A. A neural network model to control greenhouse environment[C]∥Sixth MexicanInternational Conference on Artificial Intelligence-Special Session,MICAI 2007. IEEE,2007:311-318.
[5] JARAMILLO V H,OTTEWILL J R,DUDEK R,et al. Condition monitoring of distributed systems using two-stage Bayesianinference data fusion[J]. Mechanical Systems and Signal Processing,2017,87:91-110.
[6]王静.无线传感器网络温室监测系统的设计及数据融合算法的研究[D].呼和浩特:内蒙古农业大学,2014.WANG Jing. Greenhouse monitoring system design and data fusion algorithm research based on wireless sensor network[D].Huhhot:Inner Mongolia Agricultural University,2014.(in Chinese)
[7]张品,董为浩,高大冬.一种优化的贝叶斯估计多传感器数据融合方法[J].传感技术学报,2014,27(5):643-648.ZHANG Pin,DONG Weihao,GAO Dadong. An optimal method of data fusion for multi-sensors based on Bayesian estimation[J]. Chinese Journal of Sensors and Actuators,2014,27(5):643-648.(in Chinese)
[8]孙力帆,张雅媛,郑国强,等.基于D-S证据理论的智能温室环境控制决策融合方法[J/OL].农业机械学报,2018,49(1):268-275.SUN Lifan,ZHANG Yayuan,ZHENG Guoqiang,et al. Approach to decision fusion for intelligent greenhouse environmentalcontrol based on D-S evidence theory[J/OL]. Transactions of the Chinese Society for Agricultural Machinery,2018,49(1):268-275. http:∥www. j-csam. org/jcsam/ch/reader/view_abstract. aspx? flag=1&file_no=20180133&journal_id=jcsam.DOI:10. 6041/j. issn. 1000-1298. 2018. 01. 033.(in Chinese)
[9]李琼.温室监控系统中多传感器数据融合技术的研究及应用[D].银川:宁夏大学,2013.LI Qiong. Research and application of multisensor data fusion technology in the greenhouse monitoring system[D]. Yinchuan:Ningxia University,2013.(in Chinese)
[10]周观民,李荣会.基于神经网络的传感器网络数据融合技术研究[J].计算机仿真,2011,28(10):118-120,160.ZHOU Guanmin,LI Ronghui. Sensor network based on neural network data fusion technology[J]. Computer Simulation,2011,28(10):118-120,160.(in Chinese)
[11]孙凌逸,黄先祥,蔡伟,等.基于神经网络的无线传感器网络数据融合算法[J].传感技术学报,2011,24(1):122-127.SUN Lingyi,HUANG Xianxiang,CAI Wei,et al. Data aggregation of wireless sensor networks using artificial neural networks[J]. Chinese Journal of Sensors and Actuators,2011,24(1):122-127.(in Chinese)
[12]杨帆,孟翔飞,孙建红.数据融合技术在温室环境监控系统中的应用[J].农机化研究,2012,34(4):177-180.YANG Fan,MENG Xiangfei,SUN Jianhong. Applying of data fusion technology in greenhouse environment monitoring andcontrol system[J]. Agricultural Mechanization Research,2012,34(4):177-180.(in Chinese)
[13]黄小红.传感器网络数据融合技术研究及在温室控制中的应用[D].成都:电子科技大学,2009.
[14] OUTANOUTE M,LACHHAB A,ED-DAHHAK A,et al. A neural network dynamic model for temperature and relativehumidity control under greenhouse[C]∥2015 Third International Workshop on RFID And Adaptive Wireless Sensor Networks(RAWSN). IEEE,2015:6-11.
[15]王小雪.设施番茄高效生态栽培技术要点[J].南方农业,2017,11(31):61-63.
[16]王健.番茄生长发育模型研究及其专家系统设计[D].北京:北京理工大学,2015.WANG Jian. Research on growth and development model of tomato and its expert system design[D]. Beijing:Beijing Instituteof Technology,2015.(in Chinese)
[17] LIU Yinghui,DENG Genqing. Study on data fusion of wireless monitoring system for greenhouse[C]∥2015 8th InternationalConference on Intelligent Computation Technology and Automation(ICICTA). IEEE,2015:864-866.
[18] WANG X H,XU L H,WEI R H. A new fusion structure model on greenhouse environment data and a new fusion algorithm ofsunlight[C]∥2014 International Conference on Wireless Communication and Sensor Network(WCSN). IEEE,2014:418-424.
[19] CARRASQUILLA B A,CHACN R A,SOLRZANO Q M. Using IOT resources to enhance the accuracy of overdrainmeasurements in greenhouse horticulture[C]∥2016 IEEE 36th Central American and Panama Convention(CONCAPANXXXVI). IEEE,2016:1-5.
[20]赵树林,徐鹏民,吕光杰,等.数据融合算法在农业物联网信息采集中的研究与应用[J].青岛农业大学学报(自然科学版),2016,33(1):57-60,67.ZHAO Shulin,XU Pengmin,LGuangjie,et al. Research and application of data fusion algorithm in agricultural IOTinformation collection[J]. Journal of Qingdao Agricultural University(Natural Science Edition),2016,33(1):57-60,67.(in Chinese)
[21]李峰.在农业物联网中基于卡尔曼滤波算法实现系统数据的融合处理[J].农业网络信息,2014(12):13-15.LI Feng. The system data fusion in agricultural internet of things based on Kalman filtering algorithm[J]. Agricultural NetworkInformation,2014(12):13-15.(in Chinese)
[22] LECUN Y,BENGIO Y,HINTON G. Deep learning[J]. Nature,2015,521(7553):436.
[23] HINTON G E,OSINDERO S,TEH Y W. A fast learning algorithm for deep belief nets[J]. Neural Computation,2006,18(7):1527-1554.
[24] GOODFELLOW I J,BULATOV Y,IBARZ J,et al. Multi-digit number recognition from street view imagery using deepconvolutional neural networks[C]. Cornell University:ar Xiv:1312. 6082,2013.
[25] CHENG H T,KOC L,HARMSEN J,et al. Wide&deep learning for recommender systems[C]∥Proceedings of the 1stWorkshop on Deep Learning for Recommender Systems. ACM,2016:7-10.
[26]谢勇,杜建军,李永胜,等.无公害番茄无土栽培生产技术规程[J].广东农业科学,2006(12):84-87.
[2]崔琳.北方日光温室无线传感器多数据融合技术的研究[D].沈阳:沈阳农业大学,2016.CUI Lin. Study on data analysis platform of northern sunlight greenhouse's monitor and control system[D]. Shenyang:ShenyangAgricultural University,2016.(in Chinese)
[3]熊迎军,沈明霞,陆明洲,等.温室无线传感器网络系统实时数据融合算法[J].农业工程学报,2012,28(23):160-166.XIONG Yingjun,SHEN Mingxia,LU Mingzhou,et al. Algorithm of real time data fusion for greenhouse WSN system[J].Transactions of the CSAE,2012,28(23):160-166.(in Chinese)
[4] SALAZAR R,LPEZ I,ROJANO A. A neural network model to control greenhouse environment[C]∥Sixth MexicanInternational Conference on Artificial Intelligence-Special Session,MICAI 2007. IEEE,2007:311-318.
[5] JARAMILLO V H,OTTEWILL J R,DUDEK R,et al. Condition monitoring of distributed systems using two-stage Bayesianinference data fusion[J]. Mechanical Systems and Signal Processing,2017,87:91-110.
[6]王静.无线传感器网络温室监测系统的设计及数据融合算法的研究[D].呼和浩特:内蒙古农业大学,2014.WANG Jing. Greenhouse monitoring system design and data fusion algorithm research based on wireless sensor network[D].Huhhot:Inner Mongolia Agricultural University,2014.(in Chinese)
[7]张品,董为浩,高大冬.一种优化的贝叶斯估计多传感器数据融合方法[J].传感技术学报,2014,27(5):643-648.ZHANG Pin,DONG Weihao,GAO Dadong. An optimal method of data fusion for multi-sensors based on Bayesian estimation[J]. Chinese Journal of Sensors and Actuators,2014,27(5):643-648.(in Chinese)
[8]孙力帆,张雅媛,郑国强,等.基于D-S证据理论的智能温室环境控制决策融合方法[J/OL].农业机械学报,2018,49(1):268-275.SUN Lifan,ZHANG Yayuan,ZHENG Guoqiang,et al. Approach to decision fusion for intelligent greenhouse environmentalcontrol based on D-S evidence theory[J/OL]. Transactions of the Chinese Society for Agricultural Machinery,2018,49(1):268-275. http:∥www. j-csam. org/jcsam/ch/reader/view_abstract. aspx? flag=1&file_no=20180133&journal_id=jcsam.DOI:10. 6041/j. issn. 1000-1298. 2018. 01. 033.(in Chinese)
[9]李琼.温室监控系统中多传感器数据融合技术的研究及应用[D].银川:宁夏大学,2013.LI Qiong. Research and application of multisensor data fusion technology in the greenhouse monitoring system[D]. Yinchuan:Ningxia University,2013.(in Chinese)
[10]周观民,李荣会.基于神经网络的传感器网络数据融合技术研究[J].计算机仿真,2011,28(10):118-120,160.ZHOU Guanmin,LI Ronghui. Sensor network based on neural network data fusion technology[J]. Computer Simulation,2011,28(10):118-120,160.(in Chinese)
[11]孙凌逸,黄先祥,蔡伟,等.基于神经网络的无线传感器网络数据融合算法[J].传感技术学报,2011,24(1):122-127.SUN Lingyi,HUANG Xianxiang,CAI Wei,et al. Data aggregation of wireless sensor networks using artificial neural networks[J]. Chinese Journal of Sensors and Actuators,2011,24(1):122-127.(in Chinese)
[12]杨帆,孟翔飞,孙建红.数据融合技术在温室环境监控系统中的应用[J].农机化研究,2012,34(4):177-180.YANG Fan,MENG Xiangfei,SUN Jianhong. Applying of data fusion technology in greenhouse environment monitoring andcontrol system[J]. Agricultural Mechanization Research,2012,34(4):177-180.(in Chinese)
[13]黄小红.传感器网络数据融合技术研究及在温室控制中的应用[D].成都:电子科技大学,2009.
[14] OUTANOUTE M,LACHHAB A,ED-DAHHAK A,et al. A neural network dynamic model for temperature and relativehumidity control under greenhouse[C]∥2015 Third International Workshop on RFID And Adaptive Wireless Sensor Networks(RAWSN). IEEE,2015:6-11.
[15]王小雪.设施番茄高效生态栽培技术要点[J].南方农业,2017,11(31):61-63.
[16]王健.番茄生长发育模型研究及其专家系统设计[D].北京:北京理工大学,2015.WANG Jian. Research on growth and development model of tomato and its expert system design[D]. Beijing:Beijing Instituteof Technology,2015.(in Chinese)
[17] LIU Yinghui,DENG Genqing. Study on data fusion of wireless monitoring system for greenhouse[C]∥2015 8th InternationalConference on Intelligent Computation Technology and Automation(ICICTA). IEEE,2015:864-866.
[18] WANG X H,XU L H,WEI R H. A new fusion structure model on greenhouse environment data and a new fusion algorithm ofsunlight[C]∥2014 International Conference on Wireless Communication and Sensor Network(WCSN). IEEE,2014:418-424.
[19] CARRASQUILLA B A,CHACN R A,SOLRZANO Q M. Using IOT resources to enhance the accuracy of overdrainmeasurements in greenhouse horticulture[C]∥2016 IEEE 36th Central American and Panama Convention(CONCAPANXXXVI). IEEE,2016:1-5.
[20]赵树林,徐鹏民,吕光杰,等.数据融合算法在农业物联网信息采集中的研究与应用[J].青岛农业大学学报(自然科学版),2016,33(1):57-60,67.ZHAO Shulin,XU Pengmin,LGuangjie,et al. Research and application of data fusion algorithm in agricultural IOTinformation collection[J]. Journal of Qingdao Agricultural University(Natural Science Edition),2016,33(1):57-60,67.(in Chinese)
[21]李峰.在农业物联网中基于卡尔曼滤波算法实现系统数据的融合处理[J].农业网络信息,2014(12):13-15.LI Feng. The system data fusion in agricultural internet of things based on Kalman filtering algorithm[J]. Agricultural NetworkInformation,2014(12):13-15.(in Chinese)
[22] LECUN Y,BENGIO Y,HINTON G. Deep learning[J]. Nature,2015,521(7553):436.
[23] HINTON G E,OSINDERO S,TEH Y W. A fast learning algorithm for deep belief nets[J]. Neural Computation,2006,18(7):1527-1554.
[24] GOODFELLOW I J,BULATOV Y,IBARZ J,et al. Multi-digit number recognition from street view imagery using deepconvolutional neural networks[C]. Cornell University:ar Xiv:1312. 6082,2013.
[25] CHENG H T,KOC L,HARMSEN J,et al. Wide&deep learning for recommender systems[C]∥Proceedings of the 1stWorkshop on Deep Learning for Recommender Systems. ACM,2016:7-10.
[26]谢勇,杜建军,李永胜,等.无公害番茄无土栽培生产技术规程[J].广东农业科学,2006(12):84-87.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |