Burmese Grape Leaf Disease Dataset for Computer Vision-Based Plant Health Diagnosis
Description
The Burmese Grape Leaf Disease Dataset comprises 3,103 high-quality images categorized into five distinct classes representing various conditions of grapevine leaves. This dataset is curated to support machine learning, deep learning, and computer vision-based applications for automated plant disease recognition and classification. Each image captures clear visual indicators relevant to the health status of the leaf, aiding in effective feature extraction and model training. Data Collection Details: Captured Using: 1. Realme 8 (64 MP, f/1.79 aperture) 2. Redmi Note 7 Pro Max (48 MP, f/1.79 aperture) Data Source Locations: 1. Toponer Lotkon Bagan, Kaligonj-Nagori Road, Nagarvala (Latitude: 23.88658723621705, Longitude: 90.47780500780843) 2. Itakhola Bus Stand, Narsingdi (Latitude: 23.980154076764684, Longitude: 90.7332739352483) Number of Images: 1. Healthy: 1006 2. Anthracnose (Brown Spot): 447 3. Insect Damage: 990 4. Powdery Mildew: 296 5. Leaf Spot (Yellow): 364 Data Augmentation Techniques: To enhance model generalizability and address data imbalance, the dataset was augmented using the following techniques: 1. Brightness adjustment 2. Contrast enhancement 3. Rotation (random angles) 4. Shear transformation 5. Zoom-in and zoom-out scaling Augmented Images (15,515 Images): 1. Healthy: 1006*5 = 5,030 2. Anthracnose (Brown Spot): 447*5 = 2,235 3. Insect Damage: 990*5 = 4,950 4. Powdery Mildew: 296*5= 1,480 5. Leaf Spot (Yellow): 364*5= 1,820 Key Applications: 1. Automated Disease Detection: Used to train intelligent systems capable of identifying leaf diseases in real time. 2. Precision Viticulture: Enables AI-based monitoring for better vineyard management and targeted treatment. 3. Computer Vision Research: Provides a benchmark for evaluating classification and segmentation models. 4. Transfer Learning & Mobile Deployment: Suitable for fine-tuning pre-trained CNNs and deploying lightweight models on smartphones and edge devices. 5. Explainable AI in Agriculture: Ideal for interpretability research using saliency maps and XAI tools. 6. Academic and Industrial Benchmarking: Can be used in competitions, thesis projects, or commercial AI prototypes for crop health monitoring.