Corn Kernel Counting Dataset

agriculture computer vision machine learning


Dataset associated with the March 2021 Frontiers in Robotics and AI paper "Broad Dataset and Methods for Counting and Localization of On-Ear Corn Kernels", DOI: 10.3389/frobt.2021.627009

Update Frequency



Provided in the bucket.


Managed By

Intelinair, Inc.

See all datasets managed by Intelinair, Inc..


Usage Examples


Resources on AWS

  • Description
    Folder contains the terms of use, paper, and zip file of data in COOC format. Each kernel is annotated with an instance mask and appropriate class. The datset is broken into three subcomponents. The "base" (i.e. "corn") subset contains images from a wide variety of domains such as those you might find on the internet. The "narrow" subset contains images with a single ear as would be anticipated in the application. The "many" dataset resembles the base dataset but has a huge number of ears and kernels in each image. Images were annotated for (healthy) kernels, diseased kernels, “incomplete” (i.e. not fully matured) kernels, barren tips (i.e where the top of the ear does not produce kernels), and “kernel areas” where on-ear kernels were166present but resolution/lighting prevented annotation.
    Resource type
    S3 Bucket
    Amazon Resource Name (ARN)
    AWS Region
    AWS CLI Access (No AWS account required)
    aws s3 ls s3://intelinair-data-releases/corn-kernel-counting/ --no-sign-request

Edit this dataset entry on GitHub