Sunflower & Sclerotinia Detection

Project overview

Sunflower images were annotated for a computer vision dataset built around detecting sunflower plants and identifying areas affected by Sclerotinia disease. The work centred on creating accurate bounding box annotations around visible sunflower targets and around Sclerotinia-affected areas.

The dataset was prepared as training data for an AI-based object detection model designed to detect sunflower plants and identify Sclerotinia disease in agricultural images.

Work performed

• Reviewed and organised the sunflower image dataset before annotation.
• Identified visible sunflower plants and relevant regions in each image.
• Created accurate bounding boxes around sunflower targets.
• Marked visible Sclerotinia-affected areas wherever they appeared.
• Maintained consistent labelling using the Sunflower and Sclerotinia classes.
• Carried out quality checks and corrected inaccurate or missed annotations.
• Ensured the annotations suited agricultural object detection model training.

What made it hard

• Identifying Sclerotinia-affected areas against complex agricultural backgrounds.
• Disease-affected regions were easily confused with normal plant structures.
• Small or partially visible disease symptoms required close inspection.
• Keeping bounding boxes tight around both sunflower targets and affected areas.
• Lighting conditions, plant sizes and field environments varied throughout the dataset.

Result

A high-quality agricultural detection dataset was completed covering the Sunflower and Sclerotinia classes. The finished data was prepared for training a computer vision model able to detect sunflower plants and identify Sclerotinia-affected areas.

Final deliverable

Delivered accurately annotated sunflower images with bounding boxes for Sunflower and Sclerotinia, with consistent labelling and annotation quality maintained throughout the dataset. The final output matched the project requirements and was ready for AI model training and agricultural disease detection applications.