Deep Learning for Precision Agriculture: Diagnosing Jujube Leaf Diseases via Image Classification
Description
he Jujube Leaf Disease Dataset comprises 2,819 high-quality images, categorized into four distinct classes representing different conditions of jujube leaves. This dataset is curated to support machine learning, deep learning, and computer vision applications for automated leaf disease recognition and classification. Each image was collected under natural field conditions and captures clear visual symptoms of the leaf’s health status. This ensures effective feature extraction, model training, and validation, making the dataset a valuable resource for agricultural AI applications. Data Collection Details: Captured Using: 1. Realme 8 (64 MP, f/1.79 aperture) 2. Redmi Note 10 Pro Max (48 MP, f/1.79 aperture) Images were collected from Gazipur, Bangladesh at the following locations: 1. Kalameshwar Dakshinpara, Board Bazar, Gazipur (Latitude: 23.941363, Longitude: 90.381985) 2. Kathora, Board Bazar, Gazipur (Latitude: 23.944901, Longitude: 90.365987) Number of Images: 1. Healthy 1340 2. Insect Damage 750 3. Leaf Spot 729 Total 2819 Dataset Distribution 1. Healthy Leaf – Original: 1,340 - Augmented: 3,000 2. Insect Damage – Original: 750 - Augmented: 3,000 3. Leaf Spot Disease – Original: 729 - Augmented: 3,000 Total: Original = 2,819 - Augmented = 9,000 Key Applications of the Jujube Leaf Disease Dataset Early Disease Detection – Enables AI models to identify jujube leaf diseases at an early stage, reducing crop loss. Precision Agriculture – Supports mobile or IoT-based apps for instant disease diagnosis in the field. Decision Support Systems – Assists agricultural experts and policymakers in monitoring disease spread and planning management strategies. Smart Farming Tools – Can be integrated with drones and smart cameras for large-scale field surveillance. Educational & Research Use – Acts as a benchmark dataset for deep learning, computer vision, and plant pathology research. Model Benchmarking – Useful for testing and comparing performance of different CNN architectures and transfer learning models.
Files
Institutions
- Daffodil International University