Sugarcane Diseases Detection using the Improved Grey Wolf Optimization Algorithm with Convolution Neural Network

Abstract

The Indian economy is heavily dependent on agriculture as most people have agriculture as their source of income. Therefore, Indian researchers must focus on the various challenges in this field. As there is huge diversity in the cultivation scenario in the different geographical locations of the country. However, sugarcane is one of the important crops grown in the country's major states. The major challenge other than climate change is the impact of plant diseases on the various crops. Various infections are caused by the different viruses and bacteria that lead to poor crop yield, leading to losses to farmers, which reflects losses to the country. The identification of the diseases is done by due inspection by experts in the fields; however, it is difficult for farmers to contact them, and their availability will be a challenge. Therefore, it is wise enough for the tools to be developed to extract the features from the images of the plant leaves and effectively identify the diseases. In this paper, a solution to the problem of disease identification is proposed with the help of the CNN model, the parameters of which are tuned using the grey wolf optimization neural network. The model based on IGWO is compared with the three other optimization algorithms, GWO, GA and PSO, out of which the IGWO and GA had the competitive result in accuracy comparison. However, the precision was better for PSO and GWO algorithms.