ChiliSet: A Visual Data Set of Chili Seed Varieties in Bangladesh to be used in a study and Smart Farming

This data set provides 1,394 high-resolution pictures of three popular varieties of chili seeds in Bangladesh. The photographs were taken in local markets in the period between July and August 2025 to get fresh and representative samples of seeds. The three types that were in the dataset are :
Kalo Morich (Kalo): 477 images
Bombai Morich (Bombai): 463 images
Bogurar Jhal Morich (Bogura): 454 images

The photos were taken in the RAW format with the aid of various smartphone camera models. This was so as to provide variation in the quality of the device and a natural range of lighting and color differences. The seeds were captured in varying light settings and in a range of background types to enhance dataset diversity and strength to carry out machine learning tasks.

All RAW files after collection were standardized and processed. The steps included:

Resizing images to 1024 × 1024 pixels

Converting to JPG format

Facilitating readability, uniform framing, and appearance of seed shapes and textures.

Eliminating blurred/duplicate photos.