Cabbage Detection

Project overview

Agricultural field images were annotated for a computer vision dataset focused on detecting cabbage plants in field environments. The work involved identifying visible cabbage plants and creating accurate annotations to generate high-quality training data for an AI-based object detection model.

The dataset was designed to help the model recognise cabbage plants under real-world agricultural conditions, including varying growth stages, dense vegetation and complex field backgrounds.

Work performed

• Reviewed and organised the agricultural field image dataset before annotation began.
• Identified visible cabbage plants across the field images.
• Created accurate bounding boxes around individual cabbage plants.
• Maintained consistent labelling using the Cabbage class.
• Partially visible cabbage plants were annotated wherever applicable.
• Cabbage plants were distinguished carefully from weeds, surrounding crops, soil and other vegetation.
• Plants at different growth stages and sizes were handled to the same standard.
• Quality checks were carried out and inaccurate, missing or incomplete annotations were corrected.
• Annotations were confirmed as consistent and suitable for computer vision model training.

What made it hard

• Detecting cabbage plants in dense agricultural fields.
• Weeds and other similar-looking plants were easily confused with cabbage.
• Overlapping or closely positioned cabbage plants had to be separated.
• Partially hidden plants still required annotation.
• Growth stages and plant sizes varied considerably across the dataset.
• Complex backgrounds contained soil, leaves, weeds and other crops.
• Varying lighting conditions and shadows affected plant visibility.
• Keeping bounding boxes tight around irregularly shaped cabbage plants took care.

Result

A high-quality cabbage detection dataset was completed with accurate and consistent bounding box annotations. The dataset was prepared for training a computer vision model capable of detecting cabbage plants automatically in agricultural fields.

Final deliverable

Accurate and consistent bounding box annotations for cabbage plants were delivered throughout the dataset. The final output met the project requirements and was ready for AI model training and agricultural monitoring applications.