apic.ai - Pollen color dataset

Welcome to the "Pollen color dataset" repository! This comprehensive dataset offers a collection of color-calibrated images alongside corresponding information about the plants from which the pollen originated. The dataset is compiled from pollen traps installed within beehives, capturing pollen over a specific time frame, typically ranging from 24 to 48 hours.

Dataset

The pollen color dataset class representing a dataset for each data aquisition run (e.g. subset). It can be initialized for one subset.

Available subsets:

• "2022_apic_dome": Load data for the 2022_apic_dome subset.

Each subset hat the following attributes (dictionary):

• "images": A dictionary containing the different types of illumination.
  • "white": A list of file names corresponding to white illumination.
  • "UV": A list of file names corresponding to UV illumination.
  • "IR": A list of file names corresponding to IR illumination.
• "pollen_color": Pandas DataFrame containing the pollen colors.
• "lab_results": Pandas DataFrame containing the pollen sample lab results.

How to use:

from dataset.pollen_color_dataset import PollenColorDataset

# create dataset for specific subset
dataset = PollenColorDataset("2022_apic_dome")

# access the loaded dataset
images_df = dataset.subset["images"]
pollen_color_df = dataset.subset["pollen_color"]
ondevice_color_df = datset.subset["on_device_color"]
lab_results_df = dataset.subset["lab_results"]

00_images

2022_apic_dome_images

50 samples from 4 different locations and 10 different hives are available.

00_lab-results

For each sample a dataframe describes the distribution of plants. Each row describes one sample, while the column describe the technical term of the plant. Values the respective proportion in pph.