Core Reproducibility Dataset for Practical Drone Flight-Path Selection in Stereo Camera Attitude Calibration

This dataset accompanies the manuscript “Selecting Practical Drone Flight Paths for Stereo Camera Attitude Calibration in Long-Range Landing Approach Monitoring with Sensor Priors and Observation Errors.” The dataset supports the evaluation of practical drone flight paths for calibrating two fixed ground-based cameras used for three-dimensional position measurement during landing approaches. It contains data from 30 Gazebo simulation flights: five repetitions of each calibration path P0–P4 and five repetitions of the independent validation path V0. The archive includes 18,022 synchronized stereo observation pairs, recorded YOLO bounding-box centers, ideal image projections, time-aligned Gazebo ground-truth positions, flight trajectories, camera parameters, and simulated sensor measurements. Calibration and evaluation results cover three methods: independent per-camera calibration using drone RTK, joint calibration using drone RTK, and joint calibration without drone RTK. Both yaw-only calibration and roll–pitch–yaw calibration with IMU priors are represented across different IMU accuracy levels and image observation error models. Additional files provide representative position and error time histories, flight-level cross-validation results, flight-path effort measures, and numerical controls supporting the manuscript and Supplementary Materials. Data are provided in CSV and JSON formats, with an English README, variable definitions, a mapping between manuscript figures/tables and data files, and checksums. Positions use an East–North–Up coordinate system. Position errors are referenced to time-aligned Gazebo ground truth. This release contains numerical inputs and results for independent verification. Post-processing source code, detector weights, recorded videos, and complete simulation logs are not included. Recorded videos and complete simulation logs are available from the corresponding author upon reasonable request. Please cite this dataset and the associated article, when available, when using these data.