How to get depth perseption by using two cameras


#1

I’m using the openrov development kit with two standard mobius cameras. I just don’t know how to connect the two cameras to the beagle bone black, and have the beagle bone black combine the two images to get one image with a sense of depth. Like human vision.


#2

Do the mobius camera’s have any form of live feed of the video?


#3

Hi,

I found this article about how the camera is good for live stream video.

Quote:
Loop Recording will continue recording when the card fills up by deleting
the oldest clip on the card to make room. This mode requires either a 3 or
5 min. clip length setting to work.

•Movie Flip (180 deg. rotation) will record a video right-side up when the
camera must be mounted up-side down.

•Movie Sound allows turning off the sound and eliminating a sound track
from a video file to save file space when no sound is wanted.

•Photo capture mode with frame sizes of 2304x1536, 1920x1080, or 1280x720
(no up conversion).

•Time Lapse photos with delays between images of .25*, .5*, 1, 2, 5, 10, 30
and 60 second intervals.

  • Note: The settings below 1 sec. require a photo frame size of 1920x1080
    or smaller, and they will not have a date stamp even if it is turned on.

•Recording LED toggle flashes the recording mode LED on the top and rear of
the camera.

•Movie/photo time stamp toggle

•Power on button delay (to prevent premature activation while handling)

•Auto-record when power to the camera is applied (useful for car recorder
use - no button presses needed.)

•Charging on/off toggle when plugged into a USB data port (useful for some
external USB devices)

•TV out (PAL or NTSC), which includes a playback mode with audio.

•TV display ratio (4:3 or 16:9)

•GUI program for Windows OS for easy setting of the parameters as well as
updating the firmware and the GUI program itself

Thank You,

Asim Goraya


#4

I’m not sure I saw anything regarding a live feed from that. But regardless, if you can manage to get the video out of the TV outs, you can push that to a Video decoder board and then via a USB hub, pipe that to the BeagleBone.

From there you can take a look at OpenCV tools to see if you can find libraries for processing stereo video feeds.

Good luck.