作者: Zhongli Wang , Hedong Liang , Xian Wu , Yipeng Zhao , Baigen Cai
DOI: 10.1109/ICINFA.2015.7279465
关键词: Map projection 、 Ellipse 、 Matching (graph theory) 、 Distortion (optics) 、 Computer vision 、 Projection (set theory) 、 Artificial intelligence 、 Nonlinear distortion 、 Mathematics 、 Real image 、 Transformation (function)
摘要: Fisheye camera is finding increasing number of applications in robotic vision, visual surveillance, and other fields. But the use super-wide angle fish-eye lenses causes strong distortions resulting images. A method for correction fisheye distortion using only single images linearity imaged objects presented. Spherical projection a simple effective to transform image perspective image. To this method, optic center radius projecting sphere must be known advance. The proposed algorithm based on ellipse fitting. It requires possibility extraction distorted points from straight lines 3D scene. mathematical model introduced transformation. out discussed. Contrary most former algorithms, discussed here does not depend information about real world coordinates matching points. correctness effectiveness are demonstrated with