DeepHarvest: A Hybrid Approach for Tomato Leaf Disease Severity Detection

This study presents a novel methodology for detecting the severity of tomato leaf diseases by using a hybrid model that combines Convolutional Neural Networks (CNN) and Random Forest. The model demonstrates significant performance, with an overall accuracy of 53.49%. This highlights its effectiveness in accurately classifying 10 unique severity levels. The metrics of precision, recall, and F1-scores provide valuable insights into the model's capacity to effectively detect and categorise different illness stages with accuracy. The F1-score, which is calculated as the weighted average of accuracy and recall, is 61.27. This value highlights the equilibrium that has been attained between precision and recall. By harnessing the capabilities of deep learning and ensemble approaches, this research makes a valuable contribution to the progress of plant disease detection technologies. The results highlight the possible practical uses of the model in the field of agriculture. Future research efforts may focus on further improving the accuracy of the model and enhancing its ability to perform well in real-world situations.