首页 > 分享 > 基于自适应判别深度置信网络的棉花病虫害预测

基于自适应判别深度置信网络的棉花病虫害预测

花匠小妙招
2024-09-15 16:57

摘要: 作物病虫害预测是病虫害防治的前提，利用深度学习预测作物病虫害是一个有效且具有挑战性的研究课题。该文针对深度置信网络（deep belief network，DBN）在作物病虫害预测中的训练耗时长和容易收敛于局部最优解等问题，将自适应DBN和判别限制玻尔兹曼机（restricted boltzmann machine，RBM）相结合，利用棉花生长的环境信息，提出一种基于自适应判别DBN的棉花病虫害预测模型。该模型由3层RBM网络和一个判别RBM（discriminative restricted boltzmann machine，DRBM）网络组成，通过3层RBM网络将棉花生长的环境信息数据转换到与病虫害发生相关的特征空间，通过自动学习得到层次化的特征表示，再由DRBM预测棉花病虫害的发生概率。该模型将自适应学习率引入到对比差度算法中，通过自动调整学习步长，解决了在传统DBN模型训练时学习率选择难的问题；在学习过程中通过在DRBM中引入样本的类别信息，使得训练具有类别针对性，弱化传统RBM无监督训练时易出现特征同质化问题，提高了模型的预测准确率。对实际棉花的"棉铃虫、棉蚜虫、红蜘蛛"虫害和"黄萎病、枯萎病"病害的平均预测准确率为82.840%，与传统BP神经网络模型（BPNN）、强模糊支持向量机模型（SFSVM）和RBF神经网络模型（RBFNN）分别提高19.248%，24.916%和27.774%。

关键词: 病害 / 预测 / 模型 / 棉花 / 深度置信网络 / 自适应判别

Abstract: Abstract: Cotton diseases and pests seriously affect cotton quantity. Timely and accurate prediction of diseases and pests is very important for crop growers to effectively prevent and monitor cotton diseases and pests. Cotton diseases and pests can be forecast by environmental and weather information. Through various sensors in the internet of things, it is easy to acquire a lot of environmental and weather information, and many cotton existing prediction methods, techniques and systems have been proposed. However, the occurrence and development of cotton diseases and pests involve various factors, among which there are complex interactions and mutual influences. The traditional prediction model of cotton diseases and insect pests has limited expression ability and generalization ability, and the accuracy of prediction is not high. Many existing prediction models cannot meet the actual needs of pest and disease prediction system. Therefore, the prediction of cotton diseases and pests is still a challenging problem in computer vision. In recent years, deep learning networks have won numerous contests in pattern recognition and machine learning. Deep belief network (DBN) is one of the most widely used deep learning models and has been successfully applied in many fields. DBN is a superposition model composed of several restricted Boltzmann machines (RBM). However, in DBN, there are a lot of problems, such as time-consuming to pre-train, easy to get into the local optimal solution, unsupervised training and poor generalization. An adoptive discriminant deep belief network (ADDBN) is proposed to solve the time-consuming problem in the pre-training process of DBN, and then a forecasting model of cotton diseases and pests is proposed based on environmental information and ADDBN. ADDBN is constructed by three RBMs (restricted Boltzmann machine) and a discriminant RBM (DRBM). In DRBM, the label information is introduced to training process of RBM, and the discriminant information is added into learning process through constraint on the similarity of feature vectors to improve the forecasting rate. In ADDBN, an adaptive learning rate is introduced into the contrastive divergence algorithm to accelerate the model convergence. Comparing with DBN, the proposed model has two advantages, (1) adaptive learning rate is introduced into the contrast algorithm to automatically adjust learning step, which can solve the problem to choose the learning rate in the training traditional DBN model; (2) the class information of samples is introduced into DRBM in the learning process. Then the model can be targeted trained, which can weaken the characteristic homogeneous in unsupervised training the traditional RBM and improve the forecasting accuracy of the model. Finally, a series of experiments were carried out on a dataset of cotton diseases and pests to test the performance of ADDBN. The results showed that the convergence rate is accelerated significantly and the forecasting accuracy is improved as well. The experiment results on the environmental information database of "three worms and two diseases" of cotton in recent 6 years showed that the proposed prediction model has better prediction effect than the traditional prediction model such as BPNN, SFSVM and RBFNN, the prediction performance is improved by 19.248%, 24.916% and 27.774% respectively. It is an effective method to predict crop pests and diseases with faster convergence rate, good generalization ability and higher prediction effect.

[1] 赵冰梅，李贤超，王俊刚. 2011年新疆兵团棉花病虫害发生特点及原因分析[J]. 中国棉花，2012，39(3)：9－11.Zhao Bingmei, Li Xianchao, Wang Jungang. Analysis of the characteristics and causes of cotton disease and insect pests in Xinjiang Corps in 2011[J]. Chinese cotton, 2012, 39(3): 9－11. (in Chinese with English abstract) [2] 赵冰梅，李贤超. 新疆兵团棉花中后期主要病虫害发生趋势及防治对策[J]. 中国棉花，2012，39(17)：13－14.Zhao Bingmei, Li Xianchao. Occurrence trend and control countermeasures of main cotton diseases and insect pests in middle and later stage of xinjiang production and construction crops[J]. Chinese cotton, 2012, 39(17): 13－14. (in Chinese with English abstract) [3] 叶彩玲，霍治国，丁胜利，等. 农作物病虫害气象环境成因研究进展[J]. 自然灾害学报，2005，24(1)：90－97.Ye Cailing, Huo Zhiguo, Ding Shengli, et al. Advance in study on formation of meteorological environment causing crop's diseases and insect pests[J]. Journal of Natural Disasters, 2005, 14(1): 90－97. (in Chinese with English abstract) [4] 陈怀亮，张弘，李有. 农作物病虫害发生发展气象条件及预报方法研究综述[J]. 中国农业气象，2007，28(2)： 212－216Chen Huailiang, Zhang Hong, Li You. Summary of meteorological conditions and forecasting methods of crop diseases and insect pests development[J]. Chinese Journal of Agrometeorology, 2007, 28 (2): 212－216. (in Chinese with English abstract) [5] 王翔宇，温皓杰，李鑫星，等. 农业主要病害检测与预警技术研究进展分析[J]. 农业机械学报，2016，47(9)： 266－277.Wang Xiangyu, Wen Haojie, Li Xinxing, et al. Research progress analysis of mainly agricultural diseases detection and early warning technologies[J]. Transactions of the Chinese Society for Agricultural Machinery, 2016, 47(9): 266－277. (in Chinese with English abstract) [6] 刘俊稚. 几种典型植物对大气CO2浓度升高的生理和病理响应研究[D]. 杭州：浙江大学，2010.Liu Junzhi. Study on Eco-physiological and Pathological Pesponses of Several Typical Plant Species to Elevated Atmospheric CO2 [D]. Hangzhou: Zhejiang University, 2010. (in Chinese with English abstract) [7] 石盼. 基于嵌入式的设施农作物病虫害诊断终端系统的研究与设计[D]. 银川：北方民族大学，2016.Shi Pan. The Research and Design of Amenities Crop Pests and Diseases Diagnostic Terminal System Based on Embedded Technology[D].Yinchuan: Beifang Univesity of Nationality, 2016. (in Chinese with English abstract) [8] Qinghai He, Benxue Ma, Duanyang Qu. Cotton pests and diseases detection based on image processing[J]. Telkomnika, 2013,11(6): 3445－3450. [9] Supriya S Patki, G S Sable. A review: cotton leaf disease detection [J]. IOSR Journal of VLSI and Signal Processing, 2016,6(3):78－81. [10] 赵庆展，靳光才，周文杰，等. 基于移动GIS的棉田病虫害信息采集系统[J]. 农业工程学报，2015，31(4)：183－190.Zhao Qingzhan, Jin Guangcai, Zhou Wenjie, et al. Information collection system for diseases and pests in cotton field based on mobile GIS[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2015, 31(4): 183－190. (in Chinese with English abstract) [11] 赵冰梅，李红，李贤超. 2017年新疆兵团棉花主要病虫害发生预测及防治对策[J]. 中国棉花，2017，44(4)：1－2.Zhao Bingmei, Li Hong, Li Xianchao. Prediction and prevention and control of the main diseases and insect pests of cotton in Xinjiang Corps in 2017[J]. Chinese Cotton, 2017, 44(4): 1－2. (in Chinese with English abstract) [12] 陈光绒，李小琴. 基于物联网技术的农作物病虫害自动测报系统[J]. 江苏农业科学，2015，43(4)：406－410.Chen Guang rong, Li Xiaoqin. Automatic measurement and report system of crop diseases and insect pests based on Internet of things technology[J]. Jiangsu agricultural science, 2015, 43(4): 406－410. (in Chinese with English abstract) [13] 王翔宇，温皓杰，李鑫星. 农业主要病害检测与预警技术研究进展分析[J]. 农业机械学报，2016，47(9)：266－277.Wang Xiangyu, Wen Haojie, Li Xinxing. Analysis of the research progress on the detection and early warning technology of major agricultural diseases[J]. Transaltions of the Chinese Society for agricultural machinery, 2016, 47(9): 266－277. (in Chinese with English abstract) [14] 张恩迪，张佳锐. 基于物联网的农业虫害智能监控系统[J]. 农机化研究，2015(5)：229－234.Zhang Endi, Zhang Jiarui. Intelligent monitoring system of agricultural pests based on IOT[J]. Journal of Agricultural Mechanization Research, 2015(5): 229－234. (in Chinese with English abstract) [15] Sannakki S, Rajpurohit V S, Sumira F, et al. A neural network approach for disease forecasting in grapes using weather parameters[C]// International Conference on Computing. IEEE, 2013: 1－5. [16] Mingwang Shi. Based on time series and RBF network plant disease forecasting[J]. Procedia Engineering, 2011, 15: 2384－2387. [17] Mads Dyrmann, Henrik Karstoft, Henrik Skov Midtiby. Plant species classification using deep convolutional neural network[J]. Biosystems Engineering, 2016, 151: 72－80. [18] Sladojevic S, Arsenovic M, Anderla A, et al. Deep neural networks based recognition of plant diseases by leaf image classification[J]. Computational Intelligence and Neuroscience, 2016(6): 1－11. [19] 尹文君，张大伟，严京，等. 基于深度学习的大数据空气污染预报[J]. 中国海军管理，2015(6)：46－52.Yin Wenjun, Zhang Dawei, Yan Jing, et al. Large data air pollution forecast based on deep learning [J]. Chinese Navy Management, 2015(6): 46－52. (in Chinese with English abstract) [20] 罗向龙，焦琴琴，牛力瑶，等. 基于深度学习的短时交通流预测[J]. 计算机应用研究，2017，34(1)：91－93.Luo Xianglong, Jiao Qinqin, Niu Liyao, et al. Short-term traffic flow prediction based on deep learning[J]. Application Research of Computers, 2017, 34(1): 91－93. (in Chinese with English abstract) [21] Lü Y, Duan Y, Kang W, et al. Traffic flow prediction with big data: a deep learning approach[J]. IEEE Transactions on Intelligent Transportation Systems, 2015, 16(2): 865－873. [22] Hu Q, Zhang R, Zhou Y. Transfer learning for short-term wind speed prediction with deep neural networks[J]. Renewable Energy, 2016, 85: 83－95. [23] Zhao Y, Li J, Yu L. A deep learning ensemble approach for crude oil price forecasting[J]. Energy Economics, 2017, 66: 9－16. [24] 乔俊飞，王功明，李晓理，等. 基于自适应学习率的深度信念网设计与应用[J]. 自动化学报，2017，43(8)：1339－1349Qiao Junfei, Wang Gongming, Li Xiaoli. Design and application of deep belief network with adaptive learning rate[J]. Acta Automatica Sinica, 2017, 43(8): 1339－1349. (in Chinese with English abstract) [25] Zhao Z, Jiao L, Zhao J, et al. Discriminant deep belief network for high-resolution SAR image classification[J]. Pattern Recognition, 2017, 61: 686－701. [26] 胡小平，梁承华，杨之为，等. 植物病虫害BP神经网络预测系统的研制与应用[J]. 西北农林科技大学学报(自然科学版)，2001，29(2)：73－76.Hu Xiaoping, Liang Chenghua, Yang Zhiwei, et al. Development and application of the BP neural network prediction system on plant diseases and pests[J]. Journal of Northwest Science-Technology University of Agriculture and Forest (Nat. Sci. Ed.), 2001, 29(2): 73－76. (in Chinese with English abstract) [27] 陈兵，李少昆，王克如，等. 棉花黄萎病病叶光谱特征与病情严重度的估测[J]. 中国农业科学，2007，40(12)： 2709－2715.Chen Bing, Li Shaokun, Wang Keru, et al. Spectrum characteristics of cotton single leaf infected by verticillium wilt and estimation on severity level of disease[J]. Scientia Agricultura Sinica, 2007, 40(12): 2709－2715. (in Chinese with English abstract) [28] 姜培刚. 气候变化对昌邑市农作物病虫害发生程度的影响[D]. 泰安：山东农业大学，2015.Jiang Peigang. Influence of Climate Change on the Occurrence of Crop Diseases and Insect Pests in Changyi[D].Tai'an: Shandong Agricultural University, 2015. (in Chinese with English abstract) [29] Larochelle H, Mandel M, Pascanu R, et al. Learning algorithms for the classification restricted Boltzmann machine[J]. Journal of Machine Learning Research, 2012, 13(1): 643－669. [30] 周兆永，何东健，张海辉，等. 基于深度信念网络的苹果霉心病病害程度无损检测[J]. 食品科学，2017，38(14)：297－303.Zhou Zhaoyong, He Dongjian, Zhang Haihui, et al. Non- destructive detection of moldy core in apple fruit based on deep belief network[J]. Food Science, 2017, 38(14): 297－303. (in Chinese with English abstract) [31] 杨志民，梁静，刘广利. 强模糊支持向量机在稻瘟病气象预警中的应用[J]. 中国农业大学学报，2010，15(3)： 122－128.Yang Zhimin, Liang Jing, Liu Guangli. Application of strong fuzzy support vector machine on weather early warning of rice blast[J]. Journal of China Agricultural University, 2010, 15(3): 122－128. (in Chinese with English abstract) [32] 符保龙. RBF网络在农业病虫害预测中的应用研究[J]. 安徽农业科学，2008，36(1)：388－389.Fu Baolong. Application research of RBF network in forecasting agricultural disease and insect pest[J]. Journal of Anhui Agricultural Sciences, 2008, 36(1): 388 －389. (in Chinese with English abstract)