Development of Mobile Application for Plant Disease Recognition using Convolutional Neural Network Method

Abstract

Efficient growth and improved plant yields are needed to increase profits and the economy of farmers. Plant diseases result in crop losses every year, resulting in significant losses to farmers. Therefore, it is crucial to detect plant diseases in their earlier stages. Farmers frequently seek the assistance of professionals to handle their crop problems, which is both costly and inefficient because they typically discover and identify plant diseases by visual and naked-eye observation. Hence, automatic plant disease detection is needed to improve the monitoring of large fields of crops and disease detection of plants. Numerous ways of detecting plant diseases have been developed over the years, and one of them is an applied algorithm that employs a deep learning model and is capable of detecting in particular but not limited to tomato plant diseases with a high degree of accuracy. Plant diseases are detected and classified using a convolutional neural network (CNN) architecture based on information retrieved from plant leaves such as colour, texture, and shape. After training and testing the dataset using the CNN model, the model achieved an overall 95.8% validation and training accuracy rate for the classification results. Computer vision applications are also becoming popular among researchers. Smart devices like mobile phones are readily available, affordable, and more convenient, making it more suitable to have plant disease recognition applications for mobile phones.