Research On Data Processing And Forecasting Methods Of Intelligent Plant Factory

As a high-end plant growth device,the plant growth cabinet can not only provide full-spectrum LED lighting conditions,but also provide nutritional supplies for plants without interruption for 24 hours.Compared with the cultivation method under natural conditions,it has incomparable advantages.Lettuce is one of the common vegetables in daily life and has high nutritional value.Compared with natural conditions,the plant growth cabinet can also produce high-quality lettuce.The final growth quality of lettuce is determined by various environmental factors.Under the influence of different environmental variables,the final shape and quality of lettuce will be different.When environmental variables are not suitable for lettuce growth,lettuce may grow slowly or stagnate.In view of the above problems,this article focuses on two aspects:A BP neural network model was established to explore the mapping relationship between multivariate environment fusion and basic characteristics of lettuce.First,through a single variable experiment,the mapping relationship between multiple single environmental variables and lettuce growth characteristics was summarized.Secondly,in order to explore the mapping relationship between environmental variables and lettuce growth characteristics under the fusion of multiple environmental variables.Based on the characteristics of the BP neural network model,a related model was established to explore the change trend of the growth characteristic value of lettuce under the fusion of multiple environmental variables.Finally,the accuracy of the model is verified by the test set.Build a small VGG network model to effectively identify the lettuce image characteristics in different periods.The growth process of lettuce is mainly divided into nursery stage,growing stage,growing stage and mature stage,and the image characteristics of each stage are different.In order to prevent the imbalance of the sensory system,lettuce growth is poor.This paper uses deep learning algorithms to build a small VGG network model suitable for small sample training.And realized the effective identification and classification of lettuce images in various periods.Combined with the related hardware to take regular photo recognition and classification,it can effectively prevent lettuce from growing badly.