BeanVision-3

Published: 24 August 2026| Version 1 | DOI: 10.17632/fvvf72y3gt.1
Contributor:

Description

The BeanVision-3 dataset consists of images of bean leaves in three visual conditions: Healthy, Anthracnose, and Insect. The main data set consists of 2,966 pictures. The final data set has 4508 images after augmentation. The images were taken in the field and subsequently manipulated to reduce visual noise. Each leaf was removed from its original background and was put on a simple white background. The images were all brightened for better visibility, and resized to a 3000 × 3000 pixel. The dataset is organized in class wise folders, which can be directly utilized with the common machine learning and deep learning frameworks. Healthy images were before augmentation 943 and after augmentation the size is 4,000. For Anthracnose 515 before and 4,000 after.. For insect the numbers are 1,508 and 4,000 respectively.

Files

Institutions

Categories

Computer Vision, Image Processing, Data Collection in Agriculture, Deep Learning

Licence