Optimizing Performance in Mango Plant Leaf Disease Classification through Advanced Machine Learning Techniques

Leaf diseases pose a significant threat to the productivity and quality of mango crops, necessitating effective detection and management strategies. This study presents an automated system for the detection of mango leaf diseases using machine learning techniques. Using image processing methods to extract relevant features from leaf images, various machine learning models were trained to accurately classify common mango leaf diseases. This approach involved using a comprehensive dataset of diseased and healthy mango leaves, preprocessing images, and extracting features such as color, texture, and shape. Features were extracted using MobileNetV2 and EfficientNetV2. Feature fusion was performed using a dense layer. Principal component analysis was used to reduce dimensionality. These reduced features were then fed to a support vector classifier to classify the mango leaf images. Eight different classes were considered, including the seven most common diseases in mango leaves and one class for healthy ones. The proposed model achieved a remarkable accuracy of 99.83 %. These results demonstrate that machine learning models can achieve high accuracy in the early detection of mango leaf diseases. Implementing this system in agricultural practices can significantly help farmers in timely disease management, reducing crop losses, and improving mango production.