Currently, the market for the in-ground traffic classifiers is mature, with little or no potential for growth due to the increasing
reluctance of customers such as local authorities and the Highways Agency to allow road surfaces to be disturbed. No company active in
traffic monitoring, as far as we are aware, has a video classification technology capable of providing the same quality of data as that
from in-ground detectors.
The aim of this project is to develop an image processing application that will segregate vehicles passing a camera into a number of classes,
dependent on the vehicle size and format (for example car, van, bus, HGV, articulated lorry).
We seek to create a unique product, which is a
single camera capable of both automatic number plate recognition (ANPR) and vehicle classification simultaneously. A second innovation will be
the ability of the system to classify vehicles at night without ambient illumination. A third innovation will be in the granularity of the
classification – whereas current video classification systems typically categorise vehicles into a small number classes (such as car/van/truck)
the proposed system will be capable of class schemes of up to 12 categories, offering for the first time a viable alternative to in-ground detectors.