Abstract
License plate recognition is an important practical problem with many applications in intelligent transport systems. The most important step is license plate extraction since recognition can be done by commercial character recognition systems. License plate extraction is a hard problem due to many variations in images including distance, angle, illumination, color of the plate, etc. In this paper we propose two-stage algorithm for license plate extraction. The first stage uses dynamically adjustable thresholding for determining potential license plate areas. The second stage uses vertical edge density, contrast information and entropy to extract license plate from other candidate regions. Our proposed algorithm was tested on standard benchmark images and it successfully extracted license plates.
| Original language | English |
|---|---|
| Journal | International Journal of Systems Applications, Engineering & Development |
| Publication status | Published - 2016 |
| Externally published | Yes |