Is the nano likely fast enough to run a full stero-vision system, starting from raw images down to depth map? Would the TX2 be required for this?
I’m not sure what the fast enough do you mean, also the full stero-vision system requirements, but there are many successfully projects developed at similar use cases.
You may find the interesting one from below forum as your reference: https://forums.developer.nvidia.com/c/agx-autonomous-machines/jetson-embedded-systems/jetson-projects/78
Hi, I should specify.
Basically I’m interested in using a synched camera system which will bring in 2 5mp images at 30fps. The nano will need to run the stero-vision processing pipeline to create a depth map. After that I want to run an object classifier. The goal will be to get object detection+ recognition+ location.
I have seen stero systems which pre-process the data(ex, intel real-sense) but I want to go through the process myself.