Coastal-Imaging-Research-Network / UAV-Processing-Toolbox

LEGACY - Use CIRN-Quantitative-Coastal-Imaging-Toolbox instead. Codes, documentation and discussion for the use of single stationary cameras and small drones for Argus-like sampling.
https://github.com/Coastal-Imaging-Research-Network/UAV-Processing-Toolbox/wiki/UAV-Processing-Manual
GNU General Public License v3.0
35 stars 29 forks source link

UAV-Processing-Codes (LEGACY- PLEASE USE CIRN-Quantitative-Coastal-Imaging-Toolbox)

(LEGACY- PLEASE USE CIRN-Quantitative-Coastal-Imaging-Toolbox)

Codes, documentation and discussion for the use of small multi-rotor unmanned aerial systems (UAS) for Argus-like sampling.

This repository is intended as a home for the exchange of ideas and code for the exploitation of video data from small multi-rotor unmanned aerial systems (UAS) for doing Argus-like data analysis (or any analysis). The initial repository was developed as a systematic way of dealing with image data from a Phantom III Professional collected during the 2015 BathyDuck field experiment at the US Army Corps of Engineers Field Research Facility in Sep-Oct, 2015. The experiment, as well as all of the developed methods are described in a publication (Holman et al., 2017) and in the wiki of this repository.

One of the issues in UAV data exploitation is stabilization and geolocation of collected imagery. This depends on the degree of wander and jitter in the camera location and pointing angles, things that are quite (but not completely) stable in the Phantom. The paper and code suggest one reasonable way to solve these image geometry problems for the usual cases where there is very limited ground control.

If you find bugs or have suggestions of improvements please initiate a discussion or issue on the Issues page!