KABR Drone Maneuver Test-Clip Library: 41 Wildlife Drone Clips Labeled per Frame for Autonomous Flight Policy Evaluation

Provider
Imageomics Institute (Jenna Kline)
License
CC-BY-4.0 (source)
Size
130 files
Format
CSV, MP4 (in sampled files)
Stored at
Hugging Face (https://huggingface.co/datasets/imageomics/drone-maneuver-clips)
Access
Free and public. No account with the data provider is needed.
Category
Life Sciences
Source
https://huggingface.co/datasets/imageomics/drone-maneuver-clips

About this dataset

A benchmark of 41 six-second drone video clips (180 frames at 30 fps) cut from 20 KABR (Kenyan Animal Behavior Recognition) videos recorded over 17 survey sessions at the Mpala Research Centre, Kenya, showing giraffes, plains zebras and Grevy's zebras. Each clip is indexed by the autonomous-flight maneuver it can test (launch, follow, behavior-adaptive, SoI-aware) and ships a per-frame, per-track label table with species, expert behavior label, vigilance flag, 8-class viewpoint pose where ground truth exists, bounding boxes with size class, occlusion flags and drone latitude, longitude and altitude, plus per-frame ground-truth drone actions for each maneuver. Layout: catalog/ (video_index.csv, clip_index.csv, coverage report, pose audit) and clips/ / with clip.mp4, labels.csv and maneuver_labels.csv; source videos resolve to the KABR FAIR2 release by event ID. Supports detection, tracking, behavior recognition and maneuver-conditioned evaluation of drone navigation policies. About 4.7 GB. The compilation is CC BY 4.0 by the Imageomics Institute (curated by Jenna Kline); also cite KABR and KABR-poses.

How to load this dataset in Python

import pandas as pd
df = pd.read_csv("hf://datasets/imageomics/drone-maneuver-clips/catalog/clip_index.csv")
df.head()

Data structure

File types in sampled files: CSV, MP4

Files

Example file: catalog/clip_index.csv

Analyze this dataset without downloading it

On Scigantic, KABR Drone Maneuver Test-Clip Library: 41 Wildlife Drone Clips Labeled per Frame for Autonomous Flight Policy Evaluation opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

This page is also available as Markdown and as JSON.

Related Life Sciences datasets