Annotated Mahua Leaf Dataset for Disease Detection and Precision Agriculture

The Annotated Mahua Leaf Image Dataset is a curated collection of high-resolution images of Mahua (Madhuca longifolia syn. Madhuca indica) plants captured under natural field conditions in Madhya Pradesh, India. The dataset was developed as part of the Madhya Pradesh Council of Science and Technology (MPCST), Government of Madhya Pradesh, funded research project titled "Development of an Autonomous Robotic Device for Detection, Precision Spraying, and Disease Control in Mahua Cultivation." This dataset extends our previously published raw Mahua leaf image dataset by providing manually annotated bounding-box labels for healthy and unhealthy Mahua leaves. The annotations enable the development, training, validation, and benchmarking of artificial intelligence and computer vision models for automated disease detection and object detection in precision agriculture. The images were acquired under diverse real-field conditions, including variations in illumination, background, leaf orientation, growth stages, and environmental conditions, thereby reflecting realistic agricultural scenarios. Each image has been carefully annotated using bounding boxes to identify healthy and diseased leaf regions, providing high-quality ground truth for supervised learning applications. To improve reproducibility and support robust model development, the dataset has been standardized through image preprocessing, quality verification, annotation validation, and a leakage-free train, validation, and test split. It is suitable for a wide range of applications, including object detection, disease diagnosis, computer vision, deep learning, federated learning, edge artificial intelligence, drone-assisted crop monitoring, autonomous robotic systems, and precision agriculture. The dataset is intended to serve as a benchmark resource for researchers, academicians, and developers working on intelligent agricultural systems and AI-enabled plant health monitoring.