Hello. I have a usb camera which I wish to interface with the development board in this manner…
-connect USB camera to board USB IN
-writing python application to some image processing based on opencv
-display the resulting image to HDMI out lcd display
Most PYNQ supported boards, with the default image, will ship example notebooks doing exactly what you’ve described your use case to be, e.g. the opencv facedetect notebook will take a usb webcam input and stream into directly to an HDMI output…
Hi @skalade thanks for your reply. so this would work something like…
-getting data stream via uvc usb and I can write python script on board to run and display output on hdmi lcd?
I am dragging this post back up just to ask if anyone can clarify what s going on under the hood on the board when the base overlay is loaded onto the board to perform this video streaming pipeline of USB webcam → HDMI.
I am looking to test a video stream pipeline that remains entirely in the Processing system (PS) and doesn’t cross the VDMA over to the Programmable Logic (PL). (And hopefully in the future a stream that is entirely PL based and doesn’t cross over to the PS).
If I load the board with the base overlay, and create the most simple pipeline of USB in and HDMI out, will this be entirely PS? the loading of the programmable logic with the bitstream has me confused whether it is involved.
Ideally then in the future I can introduce a video processing algorithm (also in the PS) hence would like to keep the latency to a minimum.