The Dutch National Police Services Force (KLPD) has recently received a semi-automated processing facility from Siemens Nederland. This system analyses traffic violation images, and helps the police to process more speeding offences faster with less effort. It gives the KLPD a much better facility, which requires fewer staff to operate it. The predominant input format was, and still is, photographic film rolls, which are retrieved from speed cameras monitoring the national motorway network. These cameras detect and record offences very efficiently, but their processing was inefficient and very labour-intensive. The new system aims to read from the image the radar measurement records and the number plate of the offending vehicle. It was designed to be compatible with both digital and photographic film inputs, and it incorporates both automatic and manual processing components. Photographic film is digitised by a scanner, which takes about 5s to scan a frame, thus about 40min for a typical 400-frame film roll. Optical character recognition (OCR) is used in the automatic processing task, but sometimes needs to be supplemented by manual inspection. Once the registration number has been read, it is possible to deduce its nationality, for example.
Abstract