PREDICTING PLANT GROWTH IN GREENHOUSE ENVIRONMENTS USING DEEP LEARNING

Abstract

Plant development and production forecasting are crucial tasks for greenhouse farmers and farmers in 
general. Creating designs that closely replicate growth and yield may assist growers in improving environmental 
management for higher output, healthy grant and market demand, and cheaper costs.Deep Learning (DL) and 
Machine Learning (ML) are developing as powerful new analytical tools. In controlled greenhouse circumstances, 
the proposed research combines machine learning and deep learning approaches to estimate production and plant 
development in two separate situations: tomato yield forecasting and Ficusbenjamina stem growth. In the 
prediction formulae, we use the LSTM neuron model to construct a new deep RNN. The RNN structure is utilised 
to change the intended increase parameters based on prior yield, growth, and stem diameter data, as well as 
microclimate circumstances. A comparative investigation is presented to evaluate the overall performance of the 
various solutions, which includes machine learning methods such as assist vector regression and random woody 
area regression, as well as the propose rectangle error criterion.