Description
The novel dataset is applied in plant counting utilizing Unmanned Aerial Vehicles (UAVs). The dataset includes more than 400 maize cultivars and 70,870 annotated instances captured in 306 non-overlapping images with a resolution of 5472x3648. The training set consists of 200 randomly sampled images. The images contain a variable number of tassels, ranging from 36 to 550, each manually annotated tassel with a dot.
Biological Information
Species: Maize
Latin name: Zea mays
Imaging tissue: Shoot
Function Information
Computer vision tasks: Image detection
Potential application: Plant counting
Labeled instances: 70870
Label type: Semantic segmentation
Related traits: -
Imaging
Total number of images: 306
Image format: JPEG
Average resolution: 5472×3648
Dataset size: 3.6G
Sampling geographic location: Beijing, China
Acquisition
equipment: DJI Mavic 2 Pro
Sampling platform: UAV (unmanned aerial vehicle)
Sensors: RGB (visible light imaging)
