CBL - Campus del Baix Llobregat

Projecte llegit

Títol: Drone-Based Panorama Stitching: A Study of SIFT, FLANN, and RANSAC Techniques

Estudiants que han llegit aquest projecte:

Director/a: VALERO GARCÍA, MIGUEL

Departament: DAC

Títol: Drone-Based Panorama Stitching: A Study of SIFT, FLANN, and RANSAC Techniques

Data inici oferta: 01-03-2023     Data finalització oferta: 01-11-2023


Estudis d'assignació del projecte:
    MU DRONS
Tipus: Individual
 
Lloc de realització: EETAC
 
Paraules clau:
Drone, UAV, Raspberry Pi, flight plan, panorama, stitching, Python, OpenCV, SIFT, FLANN, RANSAC, homography
 
Descripció del contingut i pla d'activitats:
 
Overview (resum en anglès):
This paper documents the tasks I accomplished during my internship and project at UPC. It provides an overview of the project's structure, objectives, and task distribution. A summary is given for the Web Application part of the project, which was handled by my teammate. This paper also details the drone and payloads used in the project and their functionalities. In the parts I was responsible for, I conducted thorough investigations and tests on the Raspberry Pi camera to obtain the best image quality during every flight test. I delved into the entire process of basic panorama stitching, encompassing features detection, descriptors matching, and transformation estimation based on the homography matrix. I compared popular feature detectors and descriptor matchers in terms of processing speed and performance, subsequently developing a panorama stitching algorithm for images captured by the drone. Finally, I provided a detailed discussion on some extra tasks that were not completed and points that could be improved upon. The paper not only stands as a detailed account of our contributions but also serves as an inspiration and a guide for future enhancements of drone-based panorama stitching.

© CBLTIC Campus del Baix Llobregat - UPC