Camera management¶
On the Camera management page of the System module, you can configure and manage your cameras for data recording, calibration, or inference, as well as create data emulators for testing inference.
Here, you can do the following:
Cameras¶
Starting with version 5.12, Robovision AI supports GenICam-compliant cameras. GenICam (Generic Interface for Cameras) is an open standard developed by the European Machine Vision Association (EMVA). It defines a common programming interface for all types of industrial cameras, regardless of their hardware interface. This standard allows Robovision to communicate with compatible cameras without custom drivers or device-specific integrations.
Note
Although GenICam is a general standard, Robovision AI currently supports only Basler cameras from the GenICam family.
Additionally, you can set up the following cameras:
- Brand: Basler
- Camera type: Area scan
- Connection type: Ethernet
- Pixel formats: RGB8 (packed), Mono8, and BayerRG8
Cameras can be later used to create camera configurations.
Camera configurations¶
To record data, perform calibration, or use your cameras for inference, you must first create a camera configuration. Camera configurations can include one or multiple cameras. When creating a new camera configuration, you will be asked to choose a configuration type. Currently, two types are available: Single camera and Multi camera. Use the multi-camera option only for algorithms that support multiview processing, for example, multiview instance segmentation. All other algorithms require a single camera configuration.
WebRTC streams¶
If you want to build a dashboard to view the output of a camera, refer to the article about setting up a WebRTC stream client.