Hello everyone, I have a rov vehicle, I use rpi4 and pixhawk 2.4.8 to control this vehicle, blueos operating system is installed on rpi4 for manual control and I can manually control the vehicle from the cockpit extension.
For a competition we will participate in, the vehicle needs to pass through a rectangular shape wirelessly and autonomously. I think it can be done with image processing, but I am not sure how to start. Can image processing and autonomous control be done via BlueOs? Where can I find resources on autonomous rov coding?
Hi @KaracaEmirhan -
Welcome to the forums!
An autonomous ROV becomes an AUV!
It is possible to accomplish your goals, several members of the forums have posted about eliminating the tether and initiating dives via WiFi with the vehicle at the surface. In most cases, ArduSub’s auto mode is used for this, and a position reference from a DVL or USBL allows the vehicle to navigate.
In your case, you would either need to simulate pilot input baed on image processing, or produce velocity estimates in the same format a position sensor would from the images you process. Using navigational tags like AprilTags or Arucodes is likely going to make things much easier, but you may find you need more computational power than the Pi 4 has available.
I’d recommend checking out pymavlink and browsing the forums to get started!
1 Like