Sunflower Head & Plant Detection

Project overview

Sunflower images were annotated for a computer vision dataset focused on detecting sunflower heads and plants in agricultural environments. Accurate bounding boxes were created around visible sunflower heads and, where a flower head was not clearly visible, around the corresponding sunflower plant.

The annotated dataset was prepared as training data for an AI-based object detection model designed to identify sunflower heads or plants in field images.

Work performed

• Reviewed and organised the sunflower image dataset before annotation.
• Identified visible sunflower heads within the field images.
• Drew accurate bounding boxes around sunflower heads.
• Annotated the sunflower plant whenever a clear flower head was not visible.
• Applied the Head_or_Plant class consistently across the dataset.
• Performed quality checks and corrected inaccurate or missed annotations.
• Ensured the bounding boxes were suitable for object detection model training.

What made it hard

• Detecting small sunflower heads among dense leaves.
• Partially visible and overlapping flower heads complicated placement.
• Annotating sunflower plants when the flower head was not clearly visible.
• Keeping bounding boxes accurate across different sizes and viewing angles.
• Complex agricultural backgrounds and varying lighting conditions.

Result

A high-quality sunflower detection dataset was completed with bounding box annotations covering sunflower heads and plants. The dataset was prepared for training an AI model capable of detecting sunflower heads or plants in agricultural field images.

Final deliverable

Accurately annotated sunflower images were delivered, using bounding boxes around visible sunflower heads or, where appropriate, around the plants themselves. The dataset maintained consistent labelling and annotation quality and was prepared for AI model training and agricultural computer vision applications.