Communications of the ACM - How the virtual inspires the real
Computer Vision: A Modern Approach
Computer Vision: A Modern Approach
Multiple View Geometry in Computer Vision
Multiple View Geometry in Computer Vision
Photo tourism: exploring photo collections in 3D
ACM SIGGRAPH 2006 Papers
Hi-index | 0.00 |
We propose the study of rectification as a foundation for the study of projective geometry, structure from motion, and vision. In particular, we argue that rectification works well as a first topic in a course on any of these subjects, an observation that has been borne out by our successful use of rectification to introduce structure from motion in a recent course. Distortion is inherent to the imaging process of a camera: angles, lengths, and relative lengths are changed. For example, parallel lines may meet in the image. rectification is the process of undoing some of these distortions, so that angles and relative lengths may be accurately measured. rectification is a compelling, intuitive and entertaining task that lends itself to implementation in a reasonable amount of time. Most importantly, rectification introduces many of the computational principles of projective geometry and structure from motion (which are important topics in computer vision), and therefore becomes an excellent testbed for learning the foundations of these subjects in a surprisingly concise manner.