Questo sito utilizza cookie di terze parti per inviarti pubblicità in linea con le tue preferenze. Se vuoi saperne di più clicca QUI 
Chiudendo questo banner, scorrendo questa pagina, cliccando su un link o proseguendo la navigazione in altra maniera, acconsenti all'uso dei cookie. OK

Image Stitching for UAV remote sensing application

I am a student of computer engineering master degree at Turin Polytechnic.
As part of the Erasmus project, for a period of six months, I have developed a thesis on Image Stitching used on a UAV at the University of Barcelona, where, at the same time I am also attending two subjects.
The kind of project should be "C mode", because the thesis is realized in another University. The task has been commissioned by Prof. Cabodi, teacher at Turin Polytechnic, and is supervised in Barcelona by Dr. Enric Pastor Llorens.

The objective of the project is to write an algorithm that is able to join top view images to create a big map. The project is done in the School of Castelldefels of UPC, within the research laboratory Icarus of EETAC Faculty.
The goal of the project is to detect an area of this map, thanks to the analysis of this images, that can be visuals or thermals.
The images are taken by the two camera aboard on an Unmanned Aerial Vehicle (UAV) built by the Icarus group leaded by Enric Pastor. first I collected information about the argument to implement a sample code.
The code written it’s about resolving the base case (join only two images) and it’s written in C using Accord.Net framework (specific for image processing).
in order to do it first some interest point in the images have been detected, using an algorithm that can extract these feature points.
After founding these points we must correlate the points in the two images to detect what points can match in the images that we have. In this project were used already existing algorithms like SIFT, SURF and RANSAC.
The first two algorithm are used to extract feature point, RANSAC is utilized to discard erroneous correlation between points and to do that exploit statistical estimation of the points correlated.
The work was based on finishing to write the code and writing the final document.

Tesi disponibile per la consultazione integrale su: http://upcommons.upc.edu
Available online: http://upcommons.upc.edu

Mostra/Nascondi contenuto.
1 Introduction The objective of the project is to write an algorithm that is able to join top view images to create a big map. The project is done in the School of Castelldefels of UPC, within the research laboratory Icarus[5]ofEETACFaculty. Thegoaloftheprojectistodetectanareaofthismap, thanksto theanalysisofthisimages,thatcanbevisualsorthermals. TheimagesaretakenbythetwocameraaboardonanUnmanned AerialVehicle(UAV)builtbytheIcarusgroupleadedbyEnricPastor[1]. First I collected information about the argument to implement a samplecode. The code written it’s about resolving the base case (join only two images) and it’s written in C using Accord.Net framework (specific forimageprocessing). inordertodoitfirstsomeinterestpointintheimageshavebeen detected,usinganalgorithmthatcanextractthesefeaturepoints. Afterfoundingthesepointswemustcorrelatethepointsinthetwo imagestodetectwhatpointscanmatchintheimagesthatwehave. InthisprojectwereusedalreadyexistingalgorithmslikeSIFT,SURF andRANSAC. Thefirsttwoalgorithmareusedtoextractfeaturepoint,RANSAC isutilizedtodiscarderroneouscorrelationbetweenpointsandtodo thatexploitstatisticalestimationofthepointscorrelated. Theworkwasbasedonfinishingtowritethecodeandwritingthe finaldocument. • Firstlythecodewillbeabletomakeabigimage,derivedbythe stitchingofmanysinglesimages. • Integration of stitching algorithm as a service in MAREA (Mid- dleware Architecture for Remote Embedded Applications) mid- dleware2.3. 5

Laurea liv.II (specialistica)

Facoltà: Ingegneria

Autore: Vincenzo Cani Contatta »

Composta da 100 pagine.

 

Questa tesi ha raggiunto 53 click dal 25/06/2013.

Disponibile in PDF, la consultazione è esclusivamente in formato digitale.