Vizualna inercijska odometrija u vanjskom prostoru

master's thesis

• Title (english): Visual inertial odometry in outdoor environments
• Author: Marko Krznarić
• Mentor: Marija Seder (mentor)
• Committee member: Marija Seder (predsjednik povjerenstva)
• Committee member: Ivan Petrović (član povjerenstva)
• Committee member: Ivan Marković (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2021-09-14, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Sustav kamere i inercijskog senzora (IMU) najbliži je elektronički ekvivalent kako živa bića vide objekte u prostoru, dok IMU predstavlja unutarnje uho koje daje osjećaj za ravnotežu u prostoru. Ovaj sustav omogućuje praćenje svog položaja u prostoru, odnosno vizualnu odometriju. Rješavanje problema vizualne odometrije samo iz monokularne slike daje skalirani pomak, budući da ne postoji način usporedbe sa stvarnim pomakom. Ovaj se problem odred̄ivanja faktora skaliranja može riješiti sustavom kalibriranih stereo kamera ili dodavanjem senzora IMU. Veliki je izazov ostvariti veću frekvenciju rada vizualne odometrije, što je moguće sa senzorom IMU koji radi na znatno višoj frekvenciji od kamere. U diplomskom je radu potrebno obraditi algoritme vizualne odometrije iz monokularne slike te riješiti problem faktora skaliranja pomaka dodavanjem senzora IMU. Razvijeni algoritam vizualne odometrije potrebno je provjeriti eksperimentalno na više različitih vanjskih prostora pod različitim uvjetima.

Abstract (english)

Camera system and an inertia sensor (IMU) is the closest electronic equivalent
to how living things see objects in space, while IMU represents the inner ear that
gives a sense of balance in space. This system allows us to monitor our position in
space, or visual odometry. Troubleshooting visual odometry only from monocular
images gives a scaled shift, since there is no way to compare with real shift. This
problem in determination of scaling factor can be solved by a system of calibrated
stereo cameras or by adding an IMU sensor. A big challenge is achieving a higher
frequency of visual odometry, which is possible with an IMU sensor that operates at a
much higher frequency than cameras. It is necessary in the graduate thesis to process
the algorithms of visual odometry from monocular images and solve the problem of
displacement scaling factor by adding an IMU sensor. The developed visual odometry
algorithm needs to be tested experimentally on different outdoor spaces under different
conditions.

• Keywords: vizualna inercijska odometrija; lokalizacija; triangulacija; epipolarna geometrija; značajke; procjena gibanja; 3D rekonstrukcija; globalna mapa; SLAM; zatvaranje petlje
• Keywords (english): visual inertial odometry; localization; triangulation; epipolar geometry; features; motion estimation; 3D reconstruction; global map; SLAM; loop closure
• Language: croatian
• URN:NBN: urn:nbn:hr:168:335421
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-13 12:49:01