Reducing optical interference, enhancing the signal, and durability are traits of PSC optical filters for ALPR.
Maximum recognition performance
ALPR (automatic license plate recognition) systems, also known as ANPR (automatic number plate recognition) systems, must work 24/7 under all light conditions and in diverse environments - from controlled indoor surroundings to harsh outdoor environments, this translates into stringent performance and durability demands. The ANPR systems apply optical character recognition (OCR), which is based on either the visible spectrum (VIS) or the near infrared (NIR) – or even a combination of these. The recognition performance is challenged by many parameters such as changing light conditions, high vehicle speed, and various number plate inks - not to mention the general performance of the applied software and hardware. So how to improve the systems recognition performance? The answer lies in enhancing the signal-to-noise ratio (SNR).
This can be achieved by choosing a good filter window. Besides being an important a part of the enclosure a filter window can function as a wavelength filter to improve the SNR. Furthermore, it can be used as an element to enhance the application design, i.e. via a dead front look hiding the optics.
Choice of base material
The choice of base material always depends on the requirements of the specific application, but materials such as PMMA, Polycarbonate and Glass are typical base materials for the filter window. However, PMMA is often the preferred choice for ANPR systems due to high optical quality, good optical filter properties, easy and cost efficient processing, price, low weight (compared to glass), UV resistance, a better scratch resistance than Polycarbonate, and finally the many possibilities to improve durability and signal quality through various surface treatments. If impact resistance is the most important requirement polycarbonate or a thicker PMMA can be considered. Filter windows of glass are of high optical quality but limitations concerning machining, impact resistance and high weight often makes the material less suitable for ANPR.
Filter windows for VIS and NIR
For VIS systems the filter window should be made of a high optical quality material and have as high VIS transmission as possible. The quality is important as small imperfections in the material or surface plane can cause optical interference. Therefore, it is important to choose clear material as optically perfect as possible, e.g. casted PMMA. The PMMA can be combined with surface treatments such as clear hard coats for increased durability or anti reflective (AR) coatings for increased VIS transmission (up to 99%).
For NIR systems the filter window should provide as high NIR transmission as possible combined with exclusion of all visible light - as the visible light creates signal noise and thereby decreased recognition performance.
A “cut-on” filter made of the casted PMMA material Solaris™ IR S306, has exactly these desired properties, and is therefore an excellent choice for ANPR applications. Besides the filter properties, a black material such as Solaris™ IR S306 provides a dead front look camouflaging the optics. For some applications this enhances the design.
Maximum NIR performance
If the performance of a standard NIR “cut-on” filter not is sufficient, it is possible to apply an AR coating which is optimised for maximum transmission in the near infrared region. The result is 99 % NIR transmission! Combined with the exclusion of all visible light this provides an extraordinary SNR, resulting in increased recognition precision and speed of the ANPR system.
Combining VIS and NIR filter
For ANPR systems which apply both NIR and VIS image technology finding a single filter window solution can be challenge. But due to the nature of PMMA it is easy to machine, which means that combining NIR and VIS filters in one single filter window often can be done by a simple machining and assembly process. In that way it is possible to have an application design with a single front window, which is optimized for both VIS and NIR. The feasibility however depends on the application specific design.