The smart sensors are used to provide the raw feature recognition and extraction capability that is a critical part of the VisionHub’s 3D triangulation methods. These features could be gauge holes on an automobile body, edges, surface shape contours or other part features of interest. These features are reported to the controller as hole centers, lines, end points, or contours. Individual 3D points or lines are computed from these features using several triangulation-based methods including classic stereo vision, structured light, and single camera modes.

Stereo Vision - Uses multiple cameras to view a common object from different angles. Each camera must also be capable of identifying a common set of geometric features of an object such as holes, edges, etc. Stereo vision uses the static relationship between the cameras to triangulate the position of the object using the common set of geometric features. If the object is shifted or rotated, the features move with it, and a new 3D location can be computed.

Structured Light - Uses at least one camera to view an object on which structured light pattern has been projected. The structured light is typically a laser stripe or stripes that are projected onto the object at an angle relative to the camera to highlight the surface shape of an object. The static relationship between the camera and the light source is used to triangulate the location of the object. If the object is shifted or rotated, the position of the structured light pattern will change relative to the camera, and a new 3D location can be computed.

Single Camera - VisionHub enables a calibrated system to operate in single camera mode during runtime operation. A single calibrated camera is sufficient to provide true 3D locations.