The Problem Statement
The customer required a cost-effective automated secure access system that can guard the entrance/exit of the large premises. The premises have two entry gates and one exit gate; two guards at each gate control the access to these entrances. The guards manually identify and note the license number of each vehicle entering/exiting the premises. The number of vehicles entering/exiting the premises at peak hours is approximately 300 per hour which made the manual system unreliable; consequently the customer wanted to replace the existing manual method with a system that can significantly reduce the time to access the gates and also keep a record of all the vehicles with the respective time tags. Moreover, the system must be integrated with the boom barrier so as to allow automatic access to known vehicles.
KritiKal employed its state-of-the-art number plate recognition system that enabled real-time identification of vehicles by capturing and analyzing a moving vehicle’s number plate using live camera feed. KritiKal’s License Plate Recognition (KLiPR) System is based on advanced pattern recognition, image processing and artificial intelligence techniques. The system uses KritiKal’s proprietary localization & segmentation algorithms for detection and OCR Engine for reading number plates in any language.
The open system architecture of KLiPR facilitates seamless integration with physical access control device (boom barrier). At the client site, each entry/exit gate is fitted with two IP Cameras, one to capture the license plate and the other to capture the driver’s face. Furthermore, high power IR lights were installed to support the cameras with night vision.
Recognized car license plates are stored into a database and can be verified against number of user-defined criteria. Whenever a vehicle enters into a facility, KLiPR captures plate images at a rate of up 30 frames per second before isolating the best picture. It then uses optical character recognition to extract the alphanumeric characters of the plate and places them into a searchable database. Once exception is incorporated into the database, the system can send out an alert whenever unauthorized plates enter the facility.
- With more than 90% accurate detection and recognition of number plate, the waiting time at the barrier reduced significantly by 80% for authorized vehicles.
- The specifically designed interactive Graphical User Interface (GUI) delivers easy control to the doorkeeper, ensuring smooth operations at the entrances.
- The Automated and Non-Intrusive system demands minimal user intervention with almost zero maintenance cost.
- 24×7 real time operation with ultra-fast video processing ensures round the clock security of the premises.
- The high storage capacity enables the user to store large amount verifiable data of approximately 4-5 months.
- The system automatically generates a daily report combining license plate image, license plate number, entry/exit time and entry/exit gate number.