Orientation and reconstruction of close-range images using lines

• Autor(a) principal: Garcia Tommaselli, Antonio Maria [UNESP]
• Data de Publicação: 2000
• Tipo de documento: Artigo de conferência
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://hdl.handle.net/11449/224719
• topic: Calibration; Computer Vision; Orientation; Reconstruction

The aim of this paper is to present two problems where straight lines were used as the primary source of control. The first problem is the orientation of images using vertical and horizontal lines for flat surface monoplotting. A mathematical model has been adapted to generate linear equations both for vertical and horizontal lines in the object space. These lines are identified and measured in the image and the rotation matrix is computed. In order to compute image scale and coordinates over the flat surface, a distance between the camera and the surface is measured using a laser device. Using these elements, the geometric elements of the flat surface, mainly areas and lengths can be accurately computed. Some experiments with real data have been performed and the results were obtained with errors within 3%. The second problem is the camera calibration both of the interior and exterior orientation parameters using straight lines. A mathematical model using straight lines that consider both interior and exterior orientation parameters as unknowns has been derived. This model is being designed to enable camera self calibration on the fly, with no control points. Straight lines that are already in the environment are to be used as control. The inner parameters being considered are the camera focal length, the coordinates of the principal point and the x-scale factor. The radial lens distortion parameters are supposed to be previously known, because they are stable and well determined with batch calibration..