Rekonstrukcija okoline zasnovana na slikama stereo kamere

undergraduate thesis

• Title (english): Environment reconstruction based on stereo camera images
• Author: Pavle Pavičić
• Mentor: Ivan Marković (mentor)
• Committee member: Ivan Marković (predsjednik povjerenstva)
• Committee member: Goran Popović (član povjerenstva)
• Committee member: Antea Hadviger (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2019-09-06, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering

Abstract

U ovom radu obrađuje se prikazivanje okoline dobivene slikama iz stereo kamere. Za što vjerniji prikaz, potrebno je odabrati adekvatan način računanja dubine slike. Kako bi to bilo ostvarivo, dubina slike se dobiva pomoću mape dispariteta koja je generirana globalnom metodom SGM. Mapa dispariteta u kombinaciji s Q matricom koja je vezana uz fizičku izvedbu kamere daju prikaz uhvaćen na slikama u 3D prostoru. Takav prikaz naziva se oblak točaka. Za spajanje više slika, odnosno oblaka točaka, u svrhu rekonstrukcije većeg prostora, potrebno je vizualnom odometrijom izračunati matrice transformacije koje sjedinjuju sve stvorene točke u jedinstven oblak točaka. Implementacija ovog algoritma ostvarena je pomoću programskih paketa i programskog jezika C++.

Abstract (english)

This thesis deals with display of the surrounding obtained by the images from the stereo camera. For a more faithful display, it is necessary to choose an adequate way of calculating the depth of the image. For this to be feasible, the depth of the image is obtained using a disparity map generated by the global SGM method. A disparity map combined with a Q matrix, which relates to the physical construction of the camera, converts a captured image in a 3D view. This type of display is called point cloud. To merge multiple point clouds in purpose of reconstructing a larger space, it is necessary to calculate the transformation matrices using visual odometry. They unify all created points into a single point cloud. Implementation of this algorithm was achieved using programming language C++, as well as C++ programming libraries.

• Keywords: stereo kamera; disparitet; oblak točaka; vizulna odometrija
• Keywords (english): stereocamera; disparity; point cloud; visual odometry
• Language: croatian
• URN:NBN: urn:nbn:hr:168:508702
• Study programme: Title: Electrical Engineering and Information Technology; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka elektrotehnike i informacijske tehnologije (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka elektrotehnike i informacijske tehnologije)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2020-01-07 21:00:27