Tomato Leaf Disease Detection and Classification using MobileNetV2 and Extreme Learning Method: A Hybrid Approach

Tomatoes are one of the most essential vegetables in the world. It is regarded as a pillar of many countries' economies. However, tomato leaf is susceptible to a variety of diseases that can reduce or eliminate production, and for this reason, early and precise detection of tomato leaf disease is very important. As a result, various deep learning methods and extreme learning machines for tomato leaf disease diagnosis have been constructed. Extreme Learning Machine (ELM), with its short learning time and excellent generalization capabilities, has been used in feature-based sickness diagnosis approaches. This paper initially proposes a hybrid method for classifying tomato leaf diseases using transfer learning and ELM. TLMV2-ELM model includes MobileNetV2 for feature extraction that generates efficient feature vectors, which are then categorized using ELM. The TLMV2-ELM method is validated using a tomato leaf dataset, and the results of the experiment show that it outperforms current approaches with 0.99% accuracy and 0.06 loss in terms of disease detection.