Using Deep Learning to Predict Plant Growth and Yield in Greenhouse\n Environments

Abstract

Effective plant growth and yield prediction is an essential task for\ngreenhouse growers and for agriculture in general. Developing models which can\neffectively model growth and yield can help growers improve the environmental\ncontrol for better production, match supply and market demand and lower costs.\nRecent developments in Machine Learning (ML) and, in particular, Deep Learning\n(DL) can provide powerful new analytical tools. The proposed study utilises ML\nand DL techniques to predict yield and plant growth variation across two\ndifferent scenarios, tomato yield forecasting and Ficus benjamina stem growth,\nin controlled greenhouse environments. We deploy a new deep recurrent neural\nnetwork (RNN), using the Long Short-Term Memory (LSTM) neuron model, in the\nprediction formulations. Both the former yield, growth and stem diameter\nvalues, as well as the microclimate conditions, are used by the RNN\narchitecture to model the targeted growth parameters. A comparative study is\npresented, using ML methods, such as support vector regression and random\nforest regression, utilising the mean square error criterion, in order to\nevaluate the performance achieved by the different methods. Very promising\nresults, based on data that have been obtained from two greenhouses, in Belgium\nand the UK, in the framework of the EU Interreg SMARTGREEN project (2017-2021),\nare presented.\n