A Study on using Convolution Neural Networks to Predict Nitrogen Deficiency in Rice Plants

Crops depend on a lot of parameters to stay healthy, providing them with only sunlight and water does not guarantee a healthy crop and a subsequent healthy yield. Nitrogen concentration is one of the said parameters that are beneficial for any plant’s growth. It’s deficiency is often regarded as a constraint on rice output globally. The purpose of this study is to develop a Convolutional Neural Network architecture to accurately predict nitrogen deficiency in rice plants. This paper starts by running seven prebuilt CNN models. Upon analysing the aforementioned pre-built models, three custom architectures are designed and constructed, the best of the architectures yielding an accuracy of 99.85%. The model's effectiveness is measured by the model’s accuracy.