Comprehensive Analysis of Different Deep Learning Models Used for Detecting Diseases in Groundnut Plant Leafs: A Systematic Review

Abstract

Groundnuts (Arachis hypogaea L.) play a vital role in global food systems and are widely cultivated across tropical and subtropical regions. Recent advances in neural networks have demonstrated remarkable potential in image-based analytical tasks, including image classification, object detection, and disease identification. Within the context of smart agriculture, these developments can be leveraged to tackle a range of challenges, such as detecting nutritional deficiencies in plant leaves, evaluating crop yield quality and quantity, identifying seeds, monitoring soil and water conditions, and enabling early disease detection. Emerging research has explored the application of deep learning techniques - particularly Convolutional Neural Network (CNN) architectures - for characterizing plant stress through digital imagery, offering significant benefits to modern agricultural practices. Given the growing relevance of CNN-based models for groundnut cultivation and disease classification, this review examines a range of deep learning architectures, with a particular focus on well-established CNN models including AlexNet, ResNet, VGG16, VGG19, U-Net, GoogLeNet, and EfficientNet, among others. The study further analyzes the most widely adopted pre-trained models applied to plant leaf disease detection and presents preliminary findings on the use of deep learning approaches for analyzing groundnut leaves.