Our vision code should be taking images via c++ code.
Why
Currently, we have a makeshift set up of taking images and saving them separately from processing them. Taking an image as part of processing would allow for more coordinated and quickly updating system, which would allow for our frame rate to easily match our processing speed. This would also make the system much neater and faster.
P.S. Image loading takes 40ms out of the total 500ms.
To-do
Fixing image load:
@oshlern (me) will look into libraries and time their sample code and pick one with the best functionality. @widakay can help with his expertise if he wishes to do so.
This needs a raspberry pi, which means we have to get the SD card first.
Links
Proposed
Our vision code should be taking images via c++ code.
Why
Currently, we have a makeshift set up of taking images and saving them separately from processing them. Taking an image as part of processing would allow for more coordinated and quickly updating system, which would allow for our frame rate to easily match our processing speed. This would also make the system much neater and faster.
P.S. Image loading takes 40ms out of the total 500ms.
To-do
Fixing image load: @oshlern (me) will look into libraries and time their sample code and pick one with the best functionality. @widakay can help with his expertise if he wishes to do so. This needs a raspberry pi, which means we have to get the SD card first. Links