Adapting ViSP Example for Robotiq Wrist Camera

[aymane1363]

Hi,

I'm working on implementing Image-Based Visual Servoing with ViSP in Universal Robot UR3e, using the example designed for RealSense cameras provided in the VISP documentation. is it possible transition from a RealSense camera to the Robotiq Wrist Camera in the context of ViSP-based visual servoing?

[fspindle]

Of course, the only thing that needs to be done is to write a wrapper on the camera SDK to be able to retrieve the images in the form of vpImage (ViSP image format) or cv::Mat (OpenCV image format). In the latter case, we provide the vpImageConvert class, which converts an OpenCV image into a ViSP image. Once this has been done, you can use the Robotiq Wrist Camera in exactly the same way as a RealSense camera.

[ER-ROUGUI]

Hello , thank you

The wrist camera doesn't have an SDK file, as mentioned by the Robotiq support team:

<<< Hi ,

Unfortunately, we do not have such tools available for distribution as we only support the usage of our wrist camera through its Robotiq URCap on UR robots.

As a side note, I do know that our wrist camera is detected as a standard webcam when connected to a PC, so you may be able to stream its video feed as one for your project.

Best regards,>>>

However, I have been able to access the wrist camera using just the IP address on my PC from this URL: http://<my_robot_ip_address>:4242/current.jpg?annotations= So I have used OpenCV (python) to access the live camera feed.

How do you think I can adapt the ViSP code to achieve the best results without encountering problems? Thank you

[fspindle]

If with OpenCV python you can get the images, it should be possible with OpenCV C++ to get the images as a cv::Mat.
If so then use vpImage::convert() to convert to a ViSP image.