Tomato Fruit Diseases Classification using MobileNetV2

The production of tomatoes is an integral part of global food system, which provides essential nutritional value and enjoyment for economic activity in many parts of the world. However, the crop itself is very susceptible to several fruit diseases that threaten both rate and quality. Such diseases often escape notice until they have developed into advanced stages causing great loss to farmers. Manual inspections are time-consuming, error-prone and ineffective for large-scale farming. Consequently, there is an increasing demand for automated, intelligent systems capable of accurately identifying, classifying and diagnosing tomato fruit diseases early on. In this research paper, a deep learning-based solution was proposed that adopts the MobileNetV2 architecture, a lightweight convolutional neural network (CNN) optimized for speed and performance. The model was trained on a curated dataset of tomato fruit images representing four common disease categories. The image data had labels such as Blossom End Rot, Sunscald, Cracking and Splitting. Data preprocessing techniques such as resizing, normalization and augmentation were employed to improve the robustness and generalization of the model. The system under discussion achieved a training accuracy of 98% and a test accuracy of 91%, indicating it is very effective in classifying tomato fruit diseases with high precision. The lightweight architecture makes it suitable for real-time agricultural applications under conditions of limited computing resources. This study proves the potential of MobileNetV2 as a scalable, easily accessible tool for detecting diseases which is conducive to smarter farming practices and better crop management.