Artificial intelligence in plant disease detection and crop management

Plant diseases remain a major threat to global agricultural productivity, with biotic stressors causing annual crop yield losses of 20–40%, thereby increasing food insecurity amid rapid population growth. Conventional disease management approaches, which rely on visual field scouting and expert diagnosis, are limited by subjectivity, delays, and poor scalability, making them inadequate for modern large-scale agriculture.

In recent decades, artificial intelligence (AI), including machine learning (ML), deep learning (DL), and computer vision, has emerged as a transformative approach in plant pathology and precision agriculture. These technologies provide automated, scalable, and high-throughput solutions for disease detection, severity assessment, and decision support. This review explores key deep learning architectures such as convolutional neural networks (CNNs), transfer learning models (ResNet, EfficientNet, VGG), object detection frameworks (YOLO, Faster R-CNN), and vision transformers (ViT). Their applications span image-based leaf diagnosis, hyperspectral and multispectral imaging, and UAV-assisted field monitoring.

Reported results show disease classification accuracies exceeding 95% under controlled conditions, while detection models achieve mean average precision (mAP) above 90% in real-field environments. Beyond detection, AI is also applied in yield prediction, smart irrigation, integrated pest management (IPM), and soil health monitoring.