Many ANPR solutions recently are based on a closed system with a specially designed ANPR camera and a box that issued a number plate to the screen or ticket. At present many VMS manufacturers have felt it is easy to insert plate recognition software into their recording platform and offer it as an addition. With a Windows PC application-based VM deleting the closed properties of the old black box solution, the whole world a new solution is open to end-users to meet various new business requirements.
The following are some guidelines for utilizing this open approach to the development of hardware and software solutions to the efficient plate.
Choose the appropriate software package-you may already have the preferred VM, if so they might have a plug-in ANPR and may function. But ask for feedback from colleagues and colleagues before you are committed. Read the trade press for testing and comparison information. Consider the manufacturer's support. How do they plan to help you with this special installation? Do they offer training and how much does it cost? What types of documentation are available and how detailed it is? Also how much is the total cost of installation, maintenance, and ownership? Will this be a long-term investment?
Understanding Features and Problem Settings - Do I need to be a programmer or is there a simple interface to change parameters and settings to improve performance? What do the final users get - can they manage the plate database easily, can they distinguish between the list that is sought and unwanted? Can they control vehicle access rights? Can they match clear videos with the actual text number plate data and can they use a number plate to find the right footage? And can they do all this without training hours and endless support calls?
Choose the right hardware - The best option is to use a special ANPR camera. But any camera that meets the required recognition rate can be suitable.
Installation and Camera Settings - ANPR systems installed well can read plates in different conditions. But if you want to reach the recognition level of 95-98%, you must accurately calculate the camera position. For the best position, the angle between the camera axis and the direction of the vehicle must be minimal.