Gima Kolmi Leaf Dataset: A Curated Image Resource for Agricultural Disease Monitoring in Ipomoea aquatica Cultivated in Bangladesh
Description
The Gima Kolmi Leaf Dataset offers high-resolution, 700+ images of Ipomoea aquatica (Gima Kolmi) leaves, captured under natural sunlight in open field conditions in Rangpur, Bangladesh. Total Dataset Size: 7,292 images across all categories, including raw, processed, and augmented data. As a widely cultivated leafy vegetable in the country, early disease detection in Gima Kolmi is vital for ensuring yield and food security. This dataset supports research in machine learning and computer vision for plant disease classification, reflecting real-world agricultural conditions in rural Bangladesh. It is well-suited for developing AI-driven tools for precision agriculture and mobile-based plant health diagnostics. Key Features of the Gima Kolmi Leaf Dataset : 1. Total Raw Images: 730 2. Processed Images: 730 3. Augmented Dataset 1: 3,645 images 4. Augmented Dataset 2: 2,187 images 5. Two Balanced Classes: • Disease • Disease Free 6. Metadata Included: • CSV file with filename, class label, source folder, and image dimensions 7. Visual Summary & Sample Image Provided Data Collection : Device Used: Realme GT Master Edition (64 MP camera) Environment: Natural daylight, sunny conditions Location: OFRD, Rangpur, Bangladesh Coordinates: 25.72046007746578, 89.26284024499962 All images were taken manually by the author under consistent lighting and background conditions. Preprocessing Details : • Resizing: All processed images resized to 224x224 pixels • Normalization: Pixel values scaled to [-1, 1] range • White background removed (optional) • Format: JPG, JPEG, PNG Data Augmentation Techniques : Augmented_1: • Horizontal_Flip • Vertical Flip • Rotation (±15°) • Brightness/Contrast Augmented_2 : • Random Zoom • Shearing • Cropping and Padding • Color Jitter • Gaussian Noise Applications : The Gima Kolmi Leaf Dataset serves as a valuable resource for a wide range of research and development efforts, particularly in: ➨ Leaf Disease Classification using machine learning and deep learning techniques. ➨ Precision Agriculture, enabling the development of decision support tools for Bangladeshi farmers. ➨ Mobile-Based Disease Detection apps for in-field diagnosis and real-time crop monitoring. ➨ AI-Powered Agricultural Robotics for automated crop health assessment. ➨ Transfer Learning & Model Benchmarking in plant pathology-related computer vision tasks. By providing labeled leaf images from real Bangladeshi farms, this dataset enables AI-based disease detection, supports early diagnosis, and promotes sustainable agriculture through practical and academic applications.