Cabbage Size Classification

Project overview

Cabbage images were annotated for a computer vision dataset focused on detecting cabbages and classifying them according to size. Every visible cabbage was enclosed with a bounding box and assigned a size-based class.

The project prepared training data for an AI model capable of detecting cabbages and classifying them as small, medium or large.

Work performed

• Reviewed and organised the cabbage image dataset.
• Identified individual cabbages in each image.
• Created accurate bounding boxes around each cabbage.
• Every cabbage was classified as small, medium or large.
• Consistent labelling was maintained across the dataset.
• Quality checks were carried out and inaccurate bounding boxes or class labels were corrected.
• Annotations were confirmed as suitable for object detection and classification model training.

What made it hard

• Distinguishing between cabbages of similar size.
• Overlapping or partially visible cabbages had to be handled individually.
• Size classification needed to stay consistent from one image to the next.
• Camera distances, angles and lighting conditions varied across the dataset.
• Creating tight bounding boxes around individual cabbages took care.

Result

A high-quality cabbage detection and size classification dataset was completed with bounding box annotations. Each cabbage was assigned an appropriate size category, providing structured training data for an AI model.

Final deliverable

Accurately annotated cabbage images were delivered with bounding boxes and size-based labels for small, medium and large cabbages. The final dataset met the project requirements and was ready for AI model training and development.