Tomato Disease Detection and Recognition Using YOLOv8

Tomatoes are one of the most vital sources of nutrition in Algeria and across the globe, as they provide essential vitamins and minerals that significantly contribute to a healthy diet, they are also a staple in numerous culinary traditions. Like any other crop, the cultivation of tomatoes has its challenges, such as dealing with a diverse array of diseases that would affect drastically reduce the yield in terms of quality and quantity. In order to combat tomato diseases, the first step would be to identify them, as early detection of a disease can lead to early treatment and reduce any damage that it would cause. In this paper, we opted for YOLOv8 as the model and trained it to detect 3 types of tomato fruit diseases, we also tried to increase the accuracy of the model by adding upon its architecture, to get more detailed feature maps from our input images, which indeed had a positive effect on our results as it increased both our precision and accuracy. In the end, YOLOv8 is also great for real-time detection tasks, so it would be great for monitoring crops constantly.