Given an input video, the system is able to identify license-plates and recognize the characters. The output is a csv where each license-plate is linked to its frame of appearance.
The following is a break down of the pipeline and the different techniques that we used.
- Generalize to any Plate color
- Adapt to different Light Conditions
- Extend Template Matching in OCR to all characters