Benefits of a Combined Medicinal Leaf Dataset: Azadirachta indica, Centella asiatica, Kalanchoe for Deep Learning Applications

This collection includes high-resolution photos of medicinal plant leaves that were collected for study on classifying diseases. The plants are Azadirachta indica, Centella asiatica, and Kalanchoe. From September 8 to October 22, 2025, a total of 988 photos were taken in Chandpur (Hajiganj), Bangladesh, and Daffodil Smart City, Bangladesh. The images were captured with a OnePlus 7T smartphone camera that has a 48 MP sensor, f/1.6 aperture, 26mm wide lens, 1/2.0" sensor, 0.8µm pixel size, PDAF, and OIS. This made sure that the images were of excellent quality and consistent. There are seven classes in the dataset: Azadirachta indica (diseased: 124 photos, healthy: 127 images), Centella asiatica (diseased: 76 images, healthy: 102 images), and Kalanchoe (asymptomatic: 159 images, chewing insect damage: 316 images, complicated foliar damage: 84 images). The original photos were taken at a size of 3000×4000 pixels. To make machine learning processing easier, they were then compressed to 512×512 pixels. There is a CSV file that has metadata such file names, species information, class labels, collection locations, and dates. This dataset is good for research in precision agriculture, computer vision, deep learning benchmarking, explainable AI, and plant disease classification.