Building 3D Models of Trees Using a 3D Camera

U ovom diplomskom radu razvijen je program koji izrađuje 3D model drveta na temelju slike snimljene 3D kamerom. Pomoću programskog jezika C++ u razvojnom okruženju Microsoft Visual Studio 2012 i korištenjem programske biblioteke Point Cloud Library (PCL) napravljena su dva algoritma, jedan učitava sliku iz datoteke i segmentira ju na valjkaste površine RANSAC metodom, te drugi koji prikazuje i sprema sliku RGB-D kamere u datoteku. Segmentacijom na valjkaste površine želi se postići prepoznavanje najdominantnijih valjkastih objekata na sceni. Razvijeni program je ispitan na RGB-D slici stabla za različite kombinacije parametara RANSAC algoritma.In this graduate thesis, a program has been developed that creates a 3D model of a tree based on images captured with a 3D camera. Two algorithm's were created using the C++ programming language in the development environment of Microsoft Visual Studio 2012 and using the programming library Point Cloud Library (PCL). One algorithm loads an image from a file and segmentes it into cylindrical surfaces using the RANSAC approach. The other algorithm creates dispalys and saves images acquired by a RGB-D camera to a file. The developed program was tested on an RGB-D image of a tree for different combinations of the applied RANSAC algorithm parameters