Strawberry Grading Dataset 🍓

Strawberry Grading Dataset 🍓

Dataset Description
This dataset contains a collection of images of strawberries sorted into four distinct quality grades: Jumbo, A, B, and C. The images were captured from strawberries harvested in Ciwidey, Bandung, Indonesia, in September 2025. This dataset is designed for machine learning tasks such as automated fruit grading and quality control.

Classes and Quality Descriptions
Jumbo: The highest quality strawberries. They are characterized by their large size, uniform shape, vibrant red color, and absence of significant defects.

Grade A: High-quality strawberries that are slightly smaller than the "Jumbo" class, with good shape, bright red color, and minimal defects.

Grade B: Medium-quality strawberries. They may have a less uniform shape, some variations in color (e.g., small white or green spots), or minor blemishes.

Grade C: The lowest quality strawberries in the dataset. They are often smaller, have irregular shapes, and may show more noticeable defects, such as bruises, significant color variations, or other imperfections.

Dataset Split and Image Count
The dataset is divided into two main parts: a training set (for model training) and a test set (for model evaluation). The distribution of images is as follows:

Training Set:

Jumbo: 27 images

Grade A: 25 images

Grade B: 14 images

Grade C: 37 images

Test Set: This set consists of 27 images, all of which contain more than one strawberry. This unique feature is intended to test the model's robustness and its ability to accurately identify and grade individual fruits within a complex, multi-instance scene, simulating real-world scenarios more effectively.

Data Source
Location: Ciwidey, Bandung, Indonesia

Harvest Date: September 2025

Potential Applications
This dataset can be used for training and evaluating machine learning models for automated strawberry grading, quality control, and agricultural research. The multi-strawberry images in the test set are particularly useful for developing robust models capable of identifying and classifying individual fruits within a larger context.