LEARNING OUTCOMES
Basic purpose of this course is to study in more depth, on both the theoretical and the practical level, all components of a photogrammetric/computer vision process which allow the fully automatic 3D object reconstruction from images. At the same time, students will have the opportunity to become familiar with state-of-the-art technologies used in photogrammetry for the collection, processing and analysis of geospatial data. Through lab exercises and projects, the course intends to encourage student initiative towards studying recent international literature as well as writing code regarding current research subject in photogrammetry and computer vision. Successful completion of this course means that students:
- Are familiar with the scientific and technological development in today’s photogrammetric practice.
- Have an in-depth insight into methods and algorithms of computer vision as now fused with conventional photogrammetric processes.
- Have a good understanding and can describe, explain and compare algorithms and techniques of sparse/dense image matching of SFM (Structure from Motion).
- Are thus capable of fully comprehending the processes of software for automatic image- based 3D scene reconstruction and use them in various contexts.
- Are in position to design, implement and apply (in small-scale projects) automated photogrammetric / computer vision procedures (regarding image orientations, camera pre- and self-calibration, 3D reconstruction); to analyze, interpret and evaluate results (regarding accuracy and reliability); and to present this in technical reports.
- Understands and can compare SLAM (Simultaneous Localization and Mapping) algorithms.
- Has adequate knowledge of algorithms for video-based optical navigation (visual odometry) and their applications in robotics.
General Competences
- Working independently
- Teamwork
- Adapting to new situations
- Criticism and self-criticism
- Search for, analysis and synthesis of data and information, with the use of the necessary technology
SYLLABUS
Elaboration in selected topics of modern photogrammetry and computer vision. Included are lectures by faculty members and invited scientists and researchers from academia and the professional field with expertise in topics of interest. The subjects refer to state-of-the-art automated processes in photogrammetry and computer vision:
- Algorithms of automatic image orientations
- Linear solutions of image orientations
- Methods for camera calibration and self-calibration
- Techniques for sparse and dense image matching
- Comparison of SFM (Structure from Motion) algorithms
- SLAM (Simultaneous Localization and Mapping) algorithms
- Video-based visual navigation (visual odometry) and its applications in robotics.
The students will also handle exercises and prepare an individual project (open-source software or coding) or an extended critical literature review on topics of the course.
STUDENT PERFORMANCE EVALUATION
Language of evaluation: Greek
Methods of Evaluation:
• Evaluation of performance in the Lab exercises.
• Oral presentation of project.
ATTACHED BIBLIOGRAPHY
1. ASPRS, 2013. Manual of Photogrammetry. 6th edition, J. Chris McGlone (editor).
2. Luhmann T., Robson S., Kyle S., Harley I., 2006. Close Range Photogrammetry: Principles, Techniques and Applications. Whittles Publishing, Scotland.
3. Szeliski R., 2010. Computer Vision: Algorithms and Applications (draft). Springer (http://szeliski.org/Book/).
4. Hartley R., Zisserman A., 2000. Multiple View Geometry in Computer Vision. Cambridge University Press.
5. Fӧrstner W., Wrobel B. P., 2016. Photogrammetric Computer Vision. Springer.
In Greek:
1. Dermanis A., 1991. Analytical Photogrammetry. Ziti Editions, Thessaloniki
2. Kraus K., 2003. Photogrammetry. Vol 1. TEE Editions, Athens.