3D camera in accordance with the stereoscopic principle and method of detecting depth maps

摘要

A 3D camera (10) in accordance with the stereoscopic principle for detecting depth maps (52) of a monitored zone 12) is set forth which has at least two camera modules (14a-b) each having an image sensor (16a-b) in mutually offset perspectives for taking two-dimensional starting images (42) as well as a stereoscopic unit (28) which is configured for the application of a stereoscopic algorithm for generating a depth map (46, 50) in that mutually associated part regions are recognized in two two-dimensional images taken within a disparity zone from offset perspectives and their distance is calculated with reference to the disparity. In this respect, a pixel binning unit (30) is provided to generate a two-dimensional intermediate image (48) of lower resolution by binning a respective plurality of pixels from a two-dimensional image (42) and a fusion unit (32) is provided which fuses at least two depth maps (46, 50) to form a common depth map (52), with at least one of the depth maps (50) being generated from intermediate images (48).

主权项

1. A 3D camera in accordance with the stereoscopic principle for detecting depth maps of a monitored zone which has at least two camera modules each having an image sensor in mutually offset perspectives for taking two-dimensional starting images as well as a stereoscopic unit which is configured for the application of a stereoscopic algorithm for generating a depth map, in that
mutually associated part zones are recognized in two two-dimensional images within a disparity zone taken from offset perspectives and their distance is calculated with reference to the disparity, further comprising a pixel binning unit to generate a two-dimensional intermediate image of lower resolution by binning a respective plurality of pixels from a two-dimensional image; further comprising a fusion unit which fuses at least two depth maps to form a common depth map, with at least one of the depth maps being generated from intermediate images.