metadata

license: cc-by-nc-4.0
task_categories:
  - object-detection
pretty_name: WildBe
size_categories:
  - 1K<n<10K
tags:
  - drone imagery
  - agriculture
  - in the wild
dataset_info:
  features:
    - name: index
      dtype: int64
    - name: image
      dtype: image
    - name: width
      dtype: int64
    - name: height
      dtype: int64
    - name: split
      dtype: string
    - name: altitude
      dtype: float64
    - name: aperture
      dtype: float64
    - name: area
      dtype: float64
    - name: date
      dtype: string
    - name: device
      dtype: string
    - name: exposure
      dtype: float64
    - name: focal
      dtype: float64
    - name: iso
      dtype: float64
    - name: latitude_deg
      dtype: float64
    - name: latitude_dir
      dtype: string
    - name: longitude_deg
      dtype: float64
    - name: longitude_dir
      dtype: string
    - name: source_image_id
      dtype: string
    - name: time
      dtype: string
    - name: labels
      sequence:
        - name: class
          dtype: int64
        - name: label
          dtype: int64
        - name: x
          dtype: float32
        - name: 'y'
          dtype: float32
        - name: width
          dtype: float32
        - name: height
          dtype: float32

Wild Berry image dataset collected in Finnish forests and peatlands using drones

Introduction

Berry picking has long-standing traditions in Finland, yet it is challenging and can potentially be dangerous. The integration of drones equipped with advanced imaging techniques represents a transformative leap forward, optimising harvests and promising sustainable practices. We propose WildBe, the first image dataset of wild berries captured in peatlands and under the canopy of Finnish forests using drones. Unlike previous and related datasets, WildBe in- cludes new varieties of berries, such as bilberries, cloudberries, lingonberries, and crowberries, captured under severe light variations and in cluttered environments. WildBe features 3,516 images, including a total of 18,468 annotated bounding boxes.

How to use: an example of visualization

import json

import numpy as np
from datasets import load_dataset
from PIL import Image, ImageDraw

# Color map for classes
classes_color_map = {
    0: (225,15,10),
    1: (40, 150, 210),
    2: (10,0,210) ,
    3: (130,5,125) ,
}

# Load the dataset
dataset = load_dataset("FBK-TeV/WildBe", split="validation")

image_bytes = dataset[50]["image"]["bytes"]
np_image = np.frombuffer(image_bytes, dtype=np.uint8)
np_image = np_image.reshape(dataset[50]["height"], dataset[50]["width"], 3)

image = Image.fromarray(np_image)
labels = json.loads(dataset[50]["labels"])

draw = ImageDraw.Draw(image)

for label in labels:
    center_x = label["x"] * dataset[50]["width"]
    center_y = label["y"] * dataset[50]["height"]
    width = label["width"] * dataset[50]["width"]
    height = label["height"] * dataset[50]["height"]
    draw.rectangle(
        [
            (center_x - width / 2, center_y - height / 2),
            (center_x + width / 2, center_y + height / 2),
        ],
        outline=classes_color_map[label["class"]],
        width=2,
    )

image.show()

ArXiv link

https://arxiv.org/abs/2405.07550

APA Citaion

Riz, L., Povoli, S., Caraffa, A., Boscaini, D., Mekhalfi, M. L., Chippendale, P., ... & Poiesi, F. (2024). Wild Berry image dataset collected in Finnish forests and peatlands using drones. arXiv preprint arXiv:2405.07550.

Bibtex

@article{riz2024wild,
  title={Wild Berry image dataset collected in Finnish forests and peatlands using drones},
  author={Riz, Luigi and Povoli, Sergio and Caraffa, Andrea and Boscaini, Davide and Mekhalfi, Mohamed Lamine and Chippendale, Paul and Turtiainen, Marjut and Partanen, Birgitta and Ballester, Laura Smith and Noguera, Francisco Blanes and others},
  journal={arXiv preprint arXiv:2405.07550},
  year={2024}
}

Acknowledgement

The FEROX project has received funding from the European Union’s Horizon Framework Programme for Research and Innovation under the Grant Agreement no 101070440 - call HORIZON-CL4-2021-DIGITAL-EMERGING-01-10: AI, Data and Robotics at work (IA).

Partners