A High-Resolution Image Dataset of Tomato (Solanum lycopersicum) Leaves for Multi-Class Disease Detection and Classification from Bangladesh

This dataset comprises high-resolution images of tomato leaves collected from Nazirpur, Pirojpur, Bangladesh, designed for multi-class disease detection and classification tasks. The dataset includes both healthy (disease-free) and diseased leaf images, representing eight distinct disease subclasses common to tomato plants.

Dataset Composition:
‣ Total images: 1,007
‣ Disease-Free images: 287
‣ Disease images: 720, distributed among the following subclasses :
① Fusarium Wilt: 28
② Blight Leaf: 53
③ Septoria Leaf Spot: 91
④ Mosaic Virus: 51
⑤ Bacterial Spot: 74
⑥ Yellow Leaf: 18
⑦ Shot Hole Disease: 141
⑧ Leaf Curl: 264

Data Collection:
Images were captured in natural agricultural fields of Nazirpur, Pirojpur, Bangladesh. The images are raw and unprocessed, with no augmentations applied. The only preprocessing performed was cropping to focus on the leaf object, ensuring clarity and consistency for downstream machine learning applications.

Purpose and Applications:
This dataset aims to facilitate the development of robust machine learning and deep learning models for accurate disease detection and classification in tomato plants. It supports research in agricultural disease management, precision farming, and computer vision-based plant pathology.

Usage Notes:
① The images are high resolution to enable detailed feature extraction.
② No artificial data augmentation or image preprocessing has been applied, preserving the natural characteristics of the leaves.