Getting distance/depth data from OAK-D Lite

JJole · Mar 20, 2023

I need distance (depth) data from OAK-D Lite. Detection I do on Jetson board using Yolo5+TensorRT, so that means: planar 2-D data (object's width and length) I already have.
As Mr. Erik suggested once earlier: I could use API's YoloSpatialDetectionNetwork , but after I looked at it, it seemed like an overkill to me because it returns bounding boxes as well.

Is there some more lighter approach using DepthAI API and C++ to return what we need (distance only or at least as less as possible beside it)?

erik · Mar 21, 2023

Hi Jole ,
If you're doing inference on host, it would likely be easiest if you just calculate spatial coords on host as well. Thoughts?
Thanks, Erik

JJole · Mar 21, 2023

Najlepša hvala Erik !!

Small thing: this Python example is using SDK as dependency? I use C++ and API would perhaps be better choice if possible?

erik · Mar 21, 2023

Hi Jole ,
The demo I linked doesn't require depthai SDK, only depthai API (python), which is 1:1 to C++ API. So you can use chatgpt to directly convert the scripts above to a (most likely) working c++ code
Thanks, Erik