A influência das injunções de orientação relativa na calibração de um sistema dual de câmeras digitais

• Autor(a) principal: Bazan, Wimerson Sanches [UNESP]
• Data de Publicação: 2009
• Outros Autores: Garcia Tommaselli, Antonio Maria [UNESP], Galo, Mauricio [UNESP], Ruy, Roberto da Silva [UNESP]
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://ojs.c3sl.ufpr.br/ojs2/index.php/bcg/article/view/15508; http://hdl.handle.net/11449/6712
• topic: Digital Camera Calibration; Multi-Camera System Calibration; Bundle Adjustment with Additional Parameters; Relative Orientation Constraints; Digital Photogrammetry

Image acquisition systems based on multi-head arrangement of digital frame cameras, such as the commercial systems DMC, UltraCam, besides others, are attractive alternatives enabling larger imaging area when compared to a single frame camera. Considering that in these systems, cameras are tightly attached to an external mount, it is assumed that relative position and orientation between cameras are stable during image acquisition and, consequently, these constraint can be included in the calibration step. This constraint is acceptable because estimates of the relative orientation (RO) parameters between cameras, from previously estimated exterior orientation parameters, present higher and significant deviations than the expected physical variations, due to error propagation. In order to solve this problem, this work presents an approach based on simultaneous calibration of two or more cameras using constraints that state that the relative rotation matrix and the distance between the cameras head are stable. Experiments with images acquired by an arrangement of two Hasselblad H2D cameras were accomplished, without and with the mentioned constraints. The experiments showed that the calibration process with RO constraints allows better results than the approach based on single camera calibration, provided that the estimation has included only images with good target distribution.