Tomato’s Disease Identification Using Machine Learning Techniques with the Potential of AR and VR Technologies for Inclusiveness

Abstract

This article uses the machine learning (ML) process of MATLAB to generate an XML file (an image trained file) and uses the XML file and computer vision to illuminate an overview of disease detection in tomatoes. After detecting diseases, the result is analyzed with different camera resolutions, light, and distance conditions. The XML file is designed with 550 images of infected tomatoes and 1100 images of fresh tomatoes using the Viola-Jones algorithm and Haar cascade like feature extraction method. As an initial prototype, a camera-mounted hardware system and a graphical user interface (GUI) can detect and recognize only spot diseases in real-time. For analyzing purposes, the system will detect spots in different lights like sunlight, white light, yellow light, red light, and green light. Then a different resolution camera and different distance of camera position will be used to detect diseases. Results of different light, camera resolution, and distances will analyze the accuracy of the system. Spot detection is effective and can easily be added to many other visible diseases that exist for animals, humans, and crops. A system has been proposed by which it will be possible to the ability to effectively diagnose, plan, visualize, and simulate diseases using virtual platforms and 3D modeling of recorded photos could be revolutionary which can be applied to the establishment of augmented reality (AR) and virtual reality (VR). However, the system is a success in terms of reliability, with adoptable false positives.