Distance and Color Invariant Automatic License Plate Recognition System

Abstract

This paper proposes a novel technique to extract and recognize the vehicle number plates from the still images/videos. Numbers from the plates are automatically extracted and then matched with the database to identify the cars owner. System has been designed to work on the standard Islamabad Computerized Number Plates (ICT). The method is simple and extremely fast, making it suitable for many real time applications. This method has no distance, color and single plate limitations. The system has two main modules (1) plate extraction module (2) OCR module. We have developed our own SCANMP algorithm for the first module. For the second module, we have used template matching and neural networks to choose the effective method from them. We have used this system on different ICT plates and we have got above 90% accuracy with 2 mega pixel. With a higher quality camera this accuracy can be increased further.