Automatic Number Plate Recognition ANDROID APPLICATION

The main goal of this project was to design a real time android application that recognizes cars in an entrance to a parking lot by their plate

Abtract

The main goal of  this project was to design a real time android application that recognizes cars in an entrance to a  parking lot by their plate.

Guidelines

  • Designing an ANPR image processing algorithm using C++ and openCV
  • Implement the algorithm on an android platform, creating its application environment and interface

The image processing algorithm

1

 

The Application

2

 

App design stages

3

 

App interface

4

5

 

Problems Encountered

  • Combination of C++, openCV (image processing) and Java (application platform) required a lot of adjustments
  • The need of a practical real time application ruled out solutions that didn’t show satisfying performance (camera too slow etc.). We had to put a lot of effort in to finding suffiecient solutions
  • openCV unfamiliarities – made us put a lot of
    time and effort over debug
  • Working with and searching for solutions in
    openCV was somewhat unstable since openCV
    environment changes and develops constantly

Summary

  • We designed a high performance algorithm that detects license plates of cars
  • We implmented that code on an android platform, creating a stable, generic android application that detects license plates on passing cars, at the entrance to a parking lot, and saves the photos on the device, for future use
  • Through this project we learned a lot about image processing, openCV and programming