Camera freezing issues with OpenROV 2.8

camera

#1

Chronic issues here. I have OpenROV 2.8 and have recently been doing some tests with it. Unfortunately, the video feed freezes every few minutes and only is rejuvenated via a restart. Even after the video feed stops, the ROV is still operable. I can’t think of anything that would be causing this - I have all of the latest software, and there have been no leaks into the electronics tube. Any help is appreciated.


#2

Hi.

To my understanding there is an ongoing issue with camera in high contrast situations pausing. This same freezing has seen with complicated grid patterns (like the tiling on the sides of pools). When the servo moves the camera up or down the image can sometimes become too complicated for it to process and it has issues.

Maybe try running some tests with different backgrounds and see if it still freezes. If it does still freeze, it could be a issue with the wire connecting to the Beaglebone.

The white wire that connects the camera to the Beaglebone has had issues that cause the signal to drop out. If the camera continues to freeze in the above situation you can run a test on the wire. If you remove the camera from the rest of the ROV and then plug it into your laptop you should be able to view the image with a standard webcam viewer. If you play around with the wire see if the video cuts out. If it does cut out then it is more than likely this connecting wire.

If the camera works without issues on a computer, than maybe your connection between the camera and the cable was loose and was simply needing to be connected better.

Hopefully some of this info will help out. Good luck with things!


#3

I have this same issue with the latest 2.8. I’m able to re-produce easily in a dark room while my ROV boots up, when I’m not connected it starts flashing the leds. In this low-light situation the flashing causes the camera to kernel panic. When I’m on the water in the daytime the camera works much better however it usually kernel panics the same way when I attempt to adjust the camera tilt. In every case, rebooting is the only way I’ve found to reset the camera. I’m still trying to figure out if there is a camera configuration setting that might lessen the frequency of these kernel panics. This seemed to happen less frequently but in a similar way on the previous version of 2.8 which had the wheezy kernel. (As opposed to this new jessie kernel).

Edit: It does seem that with *just the kernel panic sometimes it’s possible to stop and start the cockpit software from the management page and the camera comes back. However, once it starts saying “cannot seq freq” then I have to reboot. Here’s my log from dmesg https://gist.github.com/jeremyd/2a0eb7a089524c0d4ac0f1047f98c79d


#4

Actually, my particular problem has gone away now. I re-flashed the SD card. I also noticed that the SD card defaults to SXGA resolution for the camera and that I had edited it to be 1920x1080 :slight_smile: My bad. Sorry for the noise but maybe this will help others remember that re-flashing can flush out strange issues where the ROV is half-functional.