Corn Stand Count

Project overview

Agricultural field images were annotated for a computer vision dataset focused on detecting and counting individual corn plants. Each visible corn plant was identified and annotated to generate high-quality training data for an AI-based crop detection and stand count model.

The main objective was to prepare a reliable dataset for automated corn stand counting and crop population analysis in agricultural fields.

Work performed

• Reviewed and organised the agricultural field image dataset before annotation.
• Identified individual corn plants throughout the field images.
• Created accurate bounding boxes around each visible corn plant.
• Consistent labelling was maintained using the Corn class.
• Partially visible plants were annotated where applicable.
• Distinguished corn plants carefully from weeds, soil, crop residue, and surrounding vegetation.
• Ensured individual plants were separately annotated to support accurate stand counting.
• Quality checks were performed to identify missed, duplicated, or incorrectly positioned annotations.
• Corrected inaccurate annotations and maintained consistency throughout the dataset.
• Prepared the dataset for training a computer vision model capable of corn plant detection and stand counting.

What made it hard

• Counting individual corn plants in dense field environments.
• Distinguishing corn plants from weeds and surrounding vegetation.
• Handling closely positioned and overlapping plants.
• Small corn plants were difficult to detect during early growth stages.
• Annotating partially visible plants accurately.
• Complex backgrounds contained soil, crop residue, and other vegetation.
• Different plant sizes and growth stages required consistent annotation.
• Ensuring no plants were missed or counted multiple times was critical for accurate stand count results.

Result

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

Final deliverable

Accurate and consistently labelled bounding box annotations were delivered for individual corn plants throughout the dataset. The final output was prepared according to project requirements and was ready for AI model training and automated corn stand count and crop monitoring applications.