Transformation of depth maps produced by ToF cameras

In this paper we propose a depth map transformation to be used with Time-of-Flight (ToF) depth cameras. ToF cameras measure the distance along light rays while depth map representation uses the Cartesian, z-distance. Necessary correction procedure is usually provided by the camera manufacturer, however the parameters used and algorithm details remain undisclosed. We propose a simple depth map transformation based on a geometrical relations, defined by the pinhole camera model, and independently estimated camera intrinsic parameters. The transformation was evaluated using a series of ToF measurements of a flat surface. The proposed method was compared with the post-processing method implemented by the manufacturer.