Cotton leaf disease detection using convolutional neural networks

Cotton leaf diseases greatly affect the yield as well as quality of the crops and must be accurately identified at an early stage for effective control. Conventional ways of detecting diseases depends on manual inspection, which is laborious, time-consuming and susceptible to errors. This study introduces a deep learning-based approach that applies a Convolutional Neural Network (CNN) for recognition and classification of cotton leaf diseases using several characteristics including texture, color, and shape. The suggested model is trained on a data set that contains images of both healthy and sick leaves, and it is done using data augmentation and normalization to increase performance. The CNN model is evaluated by accuracy, precision, recall, and F1-score and achieves 94.3% accuracy in classification which is higher than those using traditional machine learning. The model performed very well in differentiating different diseases. The results showed that CNN-based automated detection systems have the potential to facilitate disease control and mitigate economic loss in cotton production. Future directions include increasing the dataset, model architecture improvement, and real-time deployment for field applications.