Due to a significant rise in the number of vehicles these days, the number of incidents involving stealing of vehicles, traffic rule violation and even accidents have increased. The License Plate Detection plays a significant role in all these scenarios and can also be used for numerous applications such as automatic toll collection, traffic law enforcement, parking lot access control and road traffic monitoring. This project aims at reading the vehicle registration plate so that the data extracted from the plate can be stored and used for the above mentioned applications. In This project we use connected component analysis, segmentation and machine learning for character detection and recognition.
- This project is based on the common pattern matching technique for the recognition of a single font and a fixed size character.
- We first determine the region of interest i.e the license plate using assumptions on the characteristics of the license plate. And then we use machine learning classifiers to compare common patterns for fixed size letters obtained using segmentation performed on the license plate.
- However, to achieve good performance, large amounts of samples are needed. We have used the scikit-image, an image processing package for python and scikit-learn, a machine learning package for python. The algorithm we have followed is as shown below:
Python, sci-kit learn, sci-kit image, Linux, GitHub.