Possible algorithms in openmv

I am working on a model car for which we are planning to use openmv camera. I had a few questions with respect to this

  1. What all image processing algorithms can we implement using the openmv
  2. Is multi-threading available. That is can the openmv process more requests (ex: what if I want the camera to detect a traffic light whenever it sees one and identify the color)
  3. what is the timing constraints involved? How long does the camera take to identify the object and process the information.

Sample scenarios (Are these achievable?)

  1. Can I request the camera to detect a traffic light whenever it sees one and identify the color?
  2. In a maze, can I request it to scan the QR code and exit from that outlet?
  3. Is it possible to scan a set of common images and load it in the SD such that the camera detects it whenever its seen?

Thanks in Advance!!