Description
The Global Wheat Head Detection (GWHD) is a large and diverse collection of high-resolution RGB images labeled to develop and evaluate wheat head detection methods. The dataset includes 4698 squared patches cropped from 2219 original images collected across 11 sub-datasets from different countries. The dataset has a total of 188,445 labeled wheat heads.
Biological Information
Species: Wheat
Latin name: Triticum aestivum
Imaging tissue: Ear
Function Information
Computer vision tasks: Image detection
Potential application: Detection and localization of wheat heads
Labeled instances: 188445
Label type: Boundig boxes
Related traits: -
Imaging
Total number of images: 4698
Image format: PNG
Average resolution: 1024×1024
Dataset size: 6.5G
Sampling geographic location: Japan; France
; Canada; UK;
Switzerland; China
; Australia
Acquisition
equipment: Canon PowerShot G9 X mark II,
Olympus μ850 & Sony DSC-HX90V,
Sony alpha ILCE-6000
, Sony RX0
, FLIR Chameleon3 USB3,
Prosilica GT 3300 Allied Vision,
Canon EOS 5D mark II,
Canon 550D
Sampling platform: Manned mobile platform; Handheld visible light camera in the field; Rail-based field automated gantry
Sensors: RGB (visible light imaging)
