Parrot Optimization Algorithm Integrated with YOLOv8 for Improving the Detection of Diseases in the Cotton Plant

Agricultural crops are often susceptible to various diseases, which can significantly impact yield and economic viability. Cotton, a vital cash crop, is particularly vulnerable in this regard. Timely identification and diagnosis of cotton crop infections are crucial for protecting the crops and minimizing losses. This study proposes an innovative approach that integrates the YOLOv8 model with the Parrot Optimization Algorithm (POA) to optimize disease detection in cotton plants. The POA serves as an effective optimizer in enhancing weight selection of the YOLOv8 model and thereby improving its performance accuracy. A dataset of 4760 images encompassing four classes—diseased plant, leaf plant, fresh plant, and leaf—were used to fine-tune and evaluate the model. The results demonstrate high accuracy, with an mAP@50 of 96.3%, an mAP@95 of 96.2%, and F1-scores, precisions, and recalls of 93.3%, 93.2%, and 93.4%, respectively, confirming the model's improved detection capabilities. Compared to the base YOLOv8 model and other optimization techniques, the proposed model exhibits superior performance across various measures by effectively identifying infected areas on cotton leaves.