MLD: A Multi-Class Maize Leaf Disease Image Dataset from Bangladesh (2025)

The Maize Leaf Disease Dataset consists of 1,996 images of maize leaves collected from an orchard in Jhenaidah, Bangladesh (23.54498°N, 89.15321°E) between February 02–09, 2026. The images include both healthy and diseased leaves categorized into four classes: Blight (489 images), Fall Armyworm (525 images), Gray Spot (555 images), and Healthy (427 images).

Images were captured using iPhone 13 and Samsung Galaxy S21 smartphones under natural sunlight conditions, with ambient temperatures ranging from 22°C to 26°C. Original images (HEIF and JPG formats; 4032×3024 and 2543×3390 pixels) were converted to JPG where necessary and resized to 640×480 pixels (72 dpi) for practical use.

The dataset includes labeled images suitable for multi-class classification, disease detection, and semantic segmentation tasks in precision agriculture and computer vision research.