I think this is the single best MV solution at this price-point and for it’s scalability, gratulations for realizing such a great product!
I was thinking of some projects with this system, but somehow i couldn’t find the detailed tracking specs of the camera anywhere (sorry, i may have overlooked it in the docs). so, my question is that what tracking speeds are you able to get at VGA resolution for this camera, say for a contour tracking use-case? Can the STM32F4 process at the highest FPS without being overwhelmed? If running a full x-y (pixel-by-pixel) processing algo, how many frames can be supported by this processor?
Example: if i have to track the center of a red laser dot at VGA resolution, can the processor computer and support full 30 FPS (or highest supported by the CMOS) i.e. 30 point output?
thanks!
Thank you, currently it doesn’t process VGA (max QVGA).
FPS depends on the frame size and algorithm, anything between 10 and 24.
Ibrahim, can you send VGA images to the computer and save them to the SD card?
Yes, you can stream and save full resolution JPEG images.
thanks Ibrahim! Is this limitation due to processor speed or this is imposed by the image sensor spec?
I was planning to use OpenMV for a project involving tracking of IR blobs for a security app, similar to IR lock project (IR-LOCK: infrared target tracking for Drones & DIY Robotics by IR-LOCK, LLC — Kickstarter). Speed & latency minimization is very important for my application. Expected performance is simple intensity based blob-tracking @VGA/@QVGA @60FPS. Is it possible by writing a light-weight algo?
Processing speed and memory, note the algorithm tracks multiple blobs.