BananaID: An image dataset of banana varieties and ripeness stages in Indonesia designed for classification

Type of Data: Banana images with a resolution of 384×256 pixels
Data Format: JPG
Data Contents: Banana varieties and ripeness stages.

Number of Classes:
(1) Five popular banana varieties in Indonesia: Ambon Lumut (referred to as Ambon in the dataset), Cavendish, Raja, Saba, and Mas.
(2) Four ripeness stages: Unripe, Half-ripe, Ripe, and Overripe.

Number of Images:
(1) Total original images for variety classification = 1,960, augmented to 8,750 images.
(2) Total original images for ripeness classification = 1,960, augmented to 7,000 images.

Data Distribution:
(1) Original images per banana variety:
• Ambon = 468
• Cavendish = 380
• Raja = 424
• Saba = 312
• Mas = 376

(2) Augmented images per banana variety:
• Ambon = 1,750
• Cavendish = 1,750
• Raja = 1,750
• Saba = 1,750
• Mas = 1,750

(3) Original images per ripeness stage:
• Unripe = 408
• Half-ripe = 488
• Ripe = 616
• Overripe = 448

(4) Augmented images per ripeness stage:
• Unripe = 1,750
• Half-ripe = 1,750
• Ripe = 1,750
• Overripe = 1,750

Dataset Size:
(1) Total size of the original variety dataset = 48.77 MB
(2) Total size of the augmented variety dataset = 385.40 MB
(3) Total size of the original ripeness dataset = 48.77 MB
(4) Total size of the augmented ripeness dataset = 297.89 MB

Data Acquisition Process:
Images were captured using a high-quality DSLR camera to ensure optimal detail for computer vision analysis.

Data Source Location:
Traditional markets, minimarkets, and supermarkets across various areas of Surabaya City, East Java, Indonesia, as well as harvests from farms in Tulungagung Regency, East Java, Indonesia.

Dataset Applications:
This dataset is designed to support machine learning (ML), deep learning (DL), and computer vision (CV) applications for identifying banana varieties and ripeness stages. It is also relevant for developing image-based automated systems for pre-harvest and post-harvest analysis, including variety classification and ripeness assessment.