Razvoj mobilnog 3D skenera za mapiranje prostora

master's thesis

• Title (english): Development of a mobile 3D scanner for area mapping
• Author: Ivan Strahija
• Mentor: Filip Šuligoj (mentor)
• Granter: University of Zagreb; Faculty of Mechanical Engineering and Naval Architecture; Zagreb
• Defense date and country: 2024-07-15, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Mechanical Engineering

Abstract

Sposobnost orijentacije i navigacije u nepoznatom prostoru jedan je od osnovnih problema koje je potrebno riješiti za uspješnu integraciju autonomnih sustava u ljudsku svakodnevnicu. U diplomskom radu predstavljene su i analizirane različite SLAM (Simultaneous Localization and Mapping) metode te je razvijen prijenosni vizijski sustav koji obavlja funkcije dohvaćanja podataka i izvođenja SLAM algoritama. Kao osnova vizijskog sustava korišten je NVIDIA Jetson Xavier NX Developer Kit - kompaktno računalo velike procesorske snage koje koristi CUDA jezgre za brzu obradu podataka. Za dohvaćanje slike i podataka s inercijske mjerne jedinice korištena je Intel RealSense D435i kamera. Za navedene komponente izrađeno je i kućište koje omogućava njihovu laku montažu na već postojeće robotske sustave te time predstavlja jednostavnu mogućnost proširenja njihove funkcionalnosti i povećanja stupnja autonomnosti. Predstavljeni su i osnovni algoritmi iz OpenCV biblioteke za detekciju i deskripciju značajki na fotografijama na kojima se baziraju implementirani SLAM algoritmi. Za testiranje razvijenog sustava snimljen je skup podataka u prostoru laboratorija Regionalnog centra izvrsnosti za robotske tehnologije (CRTA). Usporedbom performansi algoritama i dobivenih rezultata vidljivo je kako se integracijom vizualne odometrije s podacima koje bilježi inercijska mjerna jedinica poboljšava kvaliteta mapiranja i lokalizacije u prostoru.

Abstract (english)

The ability to orientate and navigate in an unfamiliar space is one of the fundamental problems that has to be solved if autonomous systems were to be successfully integrated into our everyday life. This thesis introduces and analyzes various SLAM (Simultaneous Localization and Mapping) methods and also presents a portable vision system for data collection and execution of SLAM algorithms. The NVIDIA Jetson Xavier NX Developer Kit - a compact, highperformance computer utilizing CUDA cores for fast data processing serves as the foundation of the vision system. The Intel RealSense D435i camera is used for image acquisition and data collection from the inertial measurement unit. A custom housing for these components has been designed, facilitating easy mounting on existing robotic systems and providing a straightforward way of enhancing their functionality and increasing their level of autonomy. Basic algorithms from the OpenCV library for feature detection and description in images, which underpin the implemented SLAM algorithms, are also presented. For testing the developed system, a dataset was recorded in the Regional Centre of Excellence for Robotic Technology (CRTA) laboratory. A comparison of the algorithms' performance and the obtained results show how integrating visual odometry with data recorded by the inertial measurement unit improves the quality of mapping and localization in the environment.

• Keywords: SLAM; NVIDIA Jetson; Intel RealSense; autonomni sustavi
• Keywords (english): SLAM; NVIDIA Jetson; Intel RealSense; autonomous systems
• Language: croatian
• URN:NBN: urn:nbn:hr:235:785948
• Study programme: Title: Mechanical Engineering; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka strojarstva (magistar/magistra inženjer/inženjerka strojarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2024-07-10 18:16:40