Ivan Tomljenović

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
E-mail:
This email address is being protected from spambots. You need JavaScript enabled to view it.
 
 
Phone: +43 - (0)662 - 8044 - 7555
Fax: +43 - (0)662 - 8044 - 7588


Scientific profile:

Ivan Tomljenović finished his master’s degree on July 15, 2011 at the University of Zagreb, Faculty of Geodesy and obtained the title of Master’s in engineering of geodesy and geoinformatics.

He was an excellent student who was given dean's award for the best students’ paper. He also worked at the faculty as a student teaching assistant (demonstrator). During his education he worked on various projects such as: Dynamic Three-dimensional Modelling of the National Park Plitvice Lakes, Barriers and Tributary Streams, Object based image analysis as a bridge between Remote Sensing and Geoinformatics, Three-dimensional laser scanning of Beničanci facility…

His interests are inthe fields of Geodesy, Geoinformatics, Hydrography, Remote sensing, Mobile GIS/WebGIS development and research, 3D terrestrial and airborne laser technology usage and research combined with Remote sensing technologies. Currently he is employed as a fully funded PhD student at the Doctoral College GIScience, and is researching in the field of applying object based image analysis to LiDAR point clouds under the supervision of Professor Thomas Blaschke.


Research Cluster
: Representation and Data Models

 

PhD Thesis Topic: Object based analysis of airborne LiDAR data through point cloud clustering and 3d-object delineation

Abstract: 

Object-based image analysis (OBIA) has become an important methodology in remote sensing and image processing. Segmentation and classification, as main parts of OBIA, have emerged as important approaches for tangible data extraction from remotely sensed imagery. With the fast development of OBIA, airborne Light Detection And Ranging (LiDAR) has emerged as a one of few new methods for remote sensing. The use of Airborne LiDAR platforms became widely available tool for fast collection of massive amounts of spatial data in form of point clouds. These point cloud are often represented as a set of vertices in a three-dimensional coordinate system. These vertices are defined by X, Y and Z coordinates, and are typically intended to be representative of the external surface of an object.

This proposal suggests a new research approach that employs the OBIA methodology (segmentation and classification) on point cloud data obtained from airborne LiDAR systems in order to extract tangible information in the form of objects. The research will also focus on the analysis of the point cloud derivatives defined as 2.5D and 2D representations (height raster, intensity raster, first/last return raster, etc.)   and a knowledge modelling.  The workload will be divided into three phases. The first phase will focus on the investigation of the existing object extraction approaches developed for the airborne LiDAR point clouds. This phase will be concluded with the publication of an extensive review of these current methods for object extraction from airborne LiDAR data. As a secondary achievement, an Internet based, crowd sourced decision making tool will be implemented.  This tool will have its roots in the scientific literature used for the generation of the review paper. The second phase will concentrate on the generation of methods and rule sets for specific object classification from an airborne LiDAR obtained point cloud and will be based on existing and newly designed approaches in the OBIA community. Some of the old segmentation and classification approaches for satellite imagery will be tested for compatibility with the airborne LiDAR point clouds and before mentioned 2.5D and 2D deliverables. The goal is to develop a new method of analysing the airborne LiDAR data in order to more effectively extract tangible information. The second major outcome of the Author’s work will be creation of the transferable rule sets for segmentation and object classification from airborne LiDAR point clouds. These rule set will be based only on the geometrical and radiometric data stored inside the point cloud and knowledge modelling. In the third and final phase, the Author will focus on testing the transferability of such developed methods across multiple scales and various data sets.

The work carried out begins with a non-classified, strip adjusted and georeferenced point cloud dataset for the area of Biberach an den Riss (Germany) and ends with the creation of segmented and classified objects. Geometric and radiometric information stored in LiDAR data will be used as the only input for the thesis work. The results obtained from this research should provide a valuable addition to already existing OBIA approaches and give a solid foundation toward the generation of fully automated, wall-to-wall, classifications.

The main strength of this proposal lies in the modelling of the existing and newly obtained knowledge into a set of simple yet powerful rule sets for object extraction, use of only airborne LiDAR point clouds for the segmentation and classification of buildings and the transferability of such an approach to other airborne LiDAR data sets regardless of the LiDAR system used for the data collection.

KEYWORDS: OBIA, LiDAR, point cloud, object extraction, point cloud modelling 

 
First Supervisor:   Prof. Dr. Thomas Blaschke   Other Supervisors:   Dr. Dirk Tiede   Jun.-Prof. Dr. Bernhard Höfle

 

Publications:

Buildings classification from airborne LiDAR point clouds through OBIA and ontology driven approach Tomljenovic, I.; Belgiu, M.; Lampoltshammer, T.J. Buildings classification from airborne LiDAR point clouds through OBIA and ontology driven approach - EGU General Assembly Conference Abstracts, 15, Vienna, Austria, 2013.

Potential and idiosyncrasy of object-based image analysis for airborne LiDAR-based building detection Tomljenovic, I.; Blaschke, T.; Höfle, B.; Tiede, D. Potential and idiosyncrasy of object-based image analysis for airborne LiDAR-based building detection. South-Eastern European Journal of Earth Observation and Geomatics 3 (2S), 2014, 517-520.

Influence of point cloud density on the results of automated Object-Based building extraction from ALS data Tomljenovic, I. & Rousell, A. Influence of point cloud density on the results of automated Object-Based building extraction from ALS data Proceedings of the AGILE'2014 International Conference on Geographic Information Science, Castellón, June, 3-6, 2014.

Influence of crisp values on the object-based data extraction procedure from LiDAR data Tomljenovic, I. & Rousell, A. Influence of crisp values on the object-based data extraction procedure from LiDAR data. EGU General Assembly Conference Abstracts, Vienna, Austria, 2014

Ontology-Based Classification of Building Types Detected from Airborne Laser Scanning Data Belgiu, M.; Tomljenovic, I.; Lampoltshammer, T.J.; Blaschke, T.; Höfle, B. Ontology-Based Classification of Building Types Detected from Airborne Laser Scanning Data. Remote Sens. 2014, 6, 1347-1366.

LIDARscapes and OBIA Blaschke, T. & Tomljenovic, I. LIDARscapes and OBIA. ASPRS 2012 Annual Conference, Sacramento, California, March 19 23, 2012

Research Areas: Object-Based Image Analysis, Light Detection And Ranging, point cloud, building classification, point cloud clustering.

Presentation Tomljenovic Ivan