It's related to ALPR that doesn't have to be real time.
The Acquisition system saves a minute long AVI file every time a motion is detected. The files are converted into images. The goal here is to post process a series of images after a motion event is triggered to first identify the ROI for automotive and secondly extract region where License plate is mounted on the car.
The images were acquired using board level USB cameras with global shutter (using M12 lens).
I believe if the plate region is reliably located for various automotive types the license plate data can be extracted. Any help in designing the most optimum algorithm to make this happen will be appreciated.
