Detection of Mango Leaf Disease using Convolution Neural Network: A Comparative Analysis

Abstract

Mango leaf diseases pose significant challenges to the agricultural industry, requiring effective and efficient detection methods to ensure the health and productivity of mango crops. This study aims to perform a comparative analysis of various Convolutional Neural Network (CNN) architectures for the detection and classification of mango leaf diseases. The dataset consists of 4,000 images, with 500 images per class. Images were preprocessed through resizing and normalization to enhance model generalization. All images in the dataset are resized to a uniform size of 224 × 224 pixels and normalized to the range [0, 1] by dividing each pixel value by 255. The dataset was split into training, validation and testing sets in an 80:10:10 ratio. The performance of several CNN models was evaluated based on accuracy, computational efficiency, and generalization ability. The DenseNet201 model outperformed others, achieving the highest average accuracy (98.10%) and proved effective for automated disease detection in mango crops. Gall Midge was identified as the most difficult disease to classify with average accuracy of 93.41%, while Cutting Weevil and Bacterial Canker showed the highest classification accuracy, 99.51% and 99.06% respectively. Computational efficiency was a key consideration, as some models, like ResNet152V2, while accurate, required substantial computational resources, limiting their feasibility in low-resource environments. This research contributes to the advancement of automated disease detection in agriculture, offering a pathway to more efficient mango leaf disease management systems.

International Journal of Statistical Sciences, Vol. 25(2), November, 2025, pp 143-157