Basler RGB-D Camera

3D depth information in true colors

Combine spatial depth data from the Basler ToF Camera with RGB data from a 2D area scan camera and the result is a 3D point cloud in the colors seen by the human eye. The advantages: better scene understanding and more precise recognition of similar objects.

Color 3D point clouds with the Basler ToF Camera

3D point cloud in false colors

The Basler ToF Camera provides 3D data as a range map or point cloud containing the x/y/z 3D coordinates for each sensor pixel. To make the evaluation user-friendly, the points are often displayed in rainbow colors (rainbow color mapping). Depth values in the near range appear red to yellow while distant values are green to blue.

3D point cloud in RGB colors

The RGB-D camera combines the depth values of the Basler ToF camera with the separately recorded color values of an RGB camera. This allows point clouds to be displayed in the colors that are actually present. This can be used to compensate for missing depth information, to perform additional classifications based on object color, or to facilitate scene understanding.

Why use RGB-D cameras?

Enables higher hit accuracy with similarly shaped objects

Better detection and classification of objects with the help of color representation.

Example:

Look at the left part of the point cloud, can you tell whether this is an apple or an orange? It is only possible to determine the difference in RGB color. Conversely, a 2D camera would not be able to tell the difference between the photo of an apple and the real apple.

For a more precise representation of scene details

The additional color information improves scene details in places where depth information is lacking.

Example:

During depalletizing, two cartons are so close together on the pallet that they are recognized as one by the 3D camera. In the higher-resolution RGB data, on the other hand, the existing gap can be detected.

For reliable assignment of scene and image segments

Deep learning can further increase reliability by integrating prior knowledge about specific scenes. This facilitates, for example, the trouble-free use of mobile robots.

Example:

The ground can be recognized without any doubt in the 3D representation when the supporting color information is used according to this assumption: asphalt gray probably belongs to the road or floor.