Automatic number plate recognition (ANPR) systems are designed for autonomous detection and recognition of number plates. They have achieved substantial attention due to its several applications such as access control, current traffic analysis during peak hours, effective law enforcement, surveillance and vehicle theft preventions but when addressing towards real world there are many challenges that needs to be addressed.
The commercially available systems generally deal with standard number plates which are comparatively easier to recognize. Conversely, it is a difficult task to identify non-standard number plates having varying sizes, fonts and patterns. Such number plates are quite frequently used in our native environment.
The problem becomes even more complex with varying illumination conditions and different orientations. We are working to design and integrate algorithms that can improve the recognition results in our dynamic environment.