MoringaLeafNet: A multi-class leaf disease dataset for precision agriculture and deep learning research

Sabit Ahamed Preanto, Tapon Paul, Abid Khan, Md Hasan Imam Bijoy

Abstract: Moringa Oleifera, which has outstanding nutritional and health benefits, is prized around the world because its leaves are rich in essential vitamins, antioxidants, and minerals that support digestion, help the immune system, and fight inflammation. Still, growing Moringa can be difficult because diseases such as Yellow Leaf, Bacterial Leaf Spot, and Cercospora Leaf Spot are hard to detect early and spread fast, leading to a lot of damage. These illnesses cause plants to make less yield, so farmers depend on pesticides and spend more, which also damages the environment and their crops. Here, we make available the MoringaLeafNet dataset, including high-quality images of leaves from the Moringa tree affected by different diseases. The images in the dataset, gathered from March to April and August to September 2025, are divided into four classes: Healthy Leaf, Yellow Leaf, Bacterial Leaf Spot, and Cercospora Leaf Spot. We collected images from Sumi Nursery in Madhupur, Tangail, Bangladesh, and Rafin Nursery in Birulia, Savar, Bangladesh, under various weather conditions. To facilitate better use in deep learning, random rotation, flipping, and brightness/contrast adjustments were performed on the data. The dataset will help develop new disease detection systems in agriculture that allow the early recognition of Moringa leaf diseases. It can also support the development of real-time diagnostic systems that provide farmers with timely insights for decision-making.

External IDs:doi:10.1016/j.dib.2025.112174