Abstract:

This report is an introduction to the Diamond3D computer vision library. The library enables fast creation of computer vision applications. Regarding the GUI and front-end of an application, there is support for image display, 3D display, mouse input to interact with images and 3D data, infrastructure for reading images from various sources, and for outputting images and derived data to disk. Regarding the main processing of an application, the architecture supports the creation of computer vision black-boxes which can be swapped between applications, allowing reuse with minimal effort. Furthermore, the library's specific goal is to enable creation of computer vision applicaitons with a focus on 3D geometry - the vision functionality includes feature detection, camera calibration, and multi-view geometry. The purpose of this report is to provide an introduction to the library. It is not a developers' guide, but a developer should begin here.