Reality capture is the use of various technical means to capture a digital 3D model representation of a subject from the real world.
Where traditional 3D modeling techniques are slow and only as accurate as the artist or designer makes them, reality capture can capture detailed, accurate digital representations in minutes.
Reality capture can be achieved by techniques such as lidar, where lasers take detailed scans of a subject, taking measurements and creating point-cloud data in an xyz axis. This point data is reconstructed into a 3D model. While accurate, lidar only captures 3D data leaving color out.
Photogrammetry is another method by which reality capture is achieved. This type of reality capture uses photographs to reconstruct a 3D image. While this method can require multiple photos to complete a model (typically with a minimum of 2), it can also capture the colors of the subject, translating them into the texture map of the model to further save time in creating assets for games. Photogrammetry can also be used to recreate motion paths of subjects.
Even 3D cameras such as Kinnect can be used for reality capture.
Reality capture can save laborious modeling of background assets and sets for 3D games and motion pictures when realism is desired. It can also be used as a starting point for stylised assets. While fast and convenient, reality capture can reduce the need for human workers.