State-of-the-Art methods for camera calibration in computer vision and photogrammetry

Abstract

The camera calibration process has been a prominent subject of research in photogrammetry and computer vision for decades. It is an essential part of numerous applications in which it directly affects accuracy and geometric quality of results. However, it does not have a single optimal solution; thus, many methods have been proposed to handle the different applications. With the increasing adoption of digital photogrammetric applications, there comes a need for robust and flexible techniques that operate autonomously. That is what camera autocalibration, or alternatively self-calibration, aims to address. Extended research has already been and is still being done in this field in such a short amount of time so that keeping up with current trends can be difficult. That is why reviews and surveys are important to aggregate all the information. This thesis is a review focused on selected multi-image autocalibration methods. In the first part, the fundamentals of the camera calibration process and the traditional calibration methods are analyzed. A selection of state-of-the-art automated traditional methods are analyzed and compared. Then, the fundamentals of camera self-calibration and its conventional methods are reviewed. Finally, a selection of state-of-the-art self-calibration methods are analyzed and compared.