The rc_cube comes with several on-board software modules, each of which corresponds to a certain functionality and can be interfaced via its respective node in the REST-API interface.
The rc_cube offers the possibility to connect multiple 3D cameras such as the rc_visard. The image data from each device is processed in a separate camera pipeline, which consists of several different software modules. The modules inside each pipeline are pipeline specific, which means that they can have different parameters for each pipeline. The modules running outside the pipelines are global and provide data for all modules in all pipelines. An overview is given in Fig. 9.
The rc_cube’s pipeline-specific software modules can be divided into
- 3D camera modules
- which acquire image pairs and compute 3D depth information such as disparity, error, and confidence images, and are also accessible via the rc_cube’s GigE Vision/GenICam interface,
- Detection modules
- which provide a variety of detection functionalities, such as grasp point computation and object detection,
- Configuration modules
- which enable the user to perform calibrations and configure the rc_cube for specific applications.
The modules that are global for all camera pipelines running on the rc_cube are the
- Database modules
- which enable the user to configure global data available to all other modules, such as load carriers, regions of interest and grippers.