Is it possible to split the imaging module and the depth computing module?

Software C++
, , , ,
1

I’m working on a project on extracting depth of a specific item in the scene which is extremely sensitive to realtime performance.

I’m wondering can it be more efficient on depth-computing if I could manage the images pushed into depth-computing module by keeping specific Mat ROI and delete the rest part according to YOLO’s result so that I can improve the operation efficiency.

Furthermore, is it possible to achieve that on Gstreamer ?

2

Hi @JanFrog
You can use the Region of Interest feature to extract depth information only from specific zones of the image:

3

Hi @Myzhar , when the ROI is set, should the depth map be calculated within this area? Or just crop the ROI from the whole resolution depth map to output

4

Hi @lin-yuxiang
Welcome to the StereoLabs community.

The depth is calculated only in the specific area.

5

Thank you for reviewing on this post.

I have another question, can the last grab be stored in memory so that I can preprocess the next frame of the image before YOLO infers the ROI ?

6

No, this is not possible. You are always one frame late.

7

@lin-yuxiang I checked with the SDK Team to confirm this, and I was wrong.
The ROI is extracted from the full depth map.
It’s not possible to calculate the depth map on a subsection of the images.

1 Like