I have an application that needs to utilize linescan video and wonder what would be needed to have the Oak-D Lite prepare the linescan by either snipping a single vertical line and sending that or converting the linescan to video frames where each frame moved the prior image left one pixel and then appended a vertical scan line from the camera on the right side. Any advice on how this can be accomplished would be appreciated.

6 months later

Hi,

Have you had any luck with the line scanning application? I just posted a near identical question before seeing yours.

-Luke

Not as yet. I found a few python apps that do rudimentary linescan with opencv which would work with Oak-D when it acts like a webcam. I'm hoping to find time this winter to see if an Oak-D app can be done which just returns the single line to reduce bandwidth and cpu overhead. But, for my application I really need optical zoom so something like a DSLR may be the solution but would be overkill for getting a single scan line. If you come up with something please let me know! Where did you post your question? I'd like to follow it too.

G

Thanks for the reply,

Re-reading your post, I’m not sure if we’re after the same thing or not. Here is the link to my post:

https://discuss.luxonis.com/d/892-line-scanning-for-2d-depth-profile

I am trying to pull X and Z spatial coordinates for a horizontal strip across a depth frame, allowing me to plot the 2D cross-section of that slice. Had some luck with shifting the roi over with each new frame, but would prefer to pull the whole set of coordinates from a single frame.

Sound anything like your application?

-Luke

(crossposting)
I have created a simple multi-ROI spatial location calculator example, see PR here. I hope this helps!
Thanks, Erik
multi_roi_slc