There is a growing demand for road traffic data (i.e. traffic volumes, speeds, lane occupancy, vehicle classification, inter-vehicles gaps, road congestion). These traffic data are required by local and central government for traffic surveillance and control, for studies of traffic management and road safety, and for the development of transport policy. Recently, image processing technology has been finding applications in such diverse industries as product manufacturing and inspection. The transport industry is no exception, and a variety of systems by image processing have been developed for traffic data collection or monitoring during the past decade. The most important theme for these traffic applications by using image processing technology is to detect a car from the picture with accuracy and basic image processing methods are used for this. However, changes of illuminations or reflectance of objects within the scene occurs regularly. So the powerful detection is expected for practical use. Under this conditions, Neutral Network (N.N.) based on a human vision system have been used for pattern recognition and a lot of good results have been reported. So we have tried to use N.N. for a new car detection system in the field of traffic engineering. This paper shows that an image processing system using N.N. at the image analysis stage in image processing procedure has a good effect for detecting of a car.
A car detection system using neural network for image processing
Kraftfahrzeugermittlungssystem benutzt das Nervennetz zur Bildverarbeitung
1992
8 Seiten, 13 Bilder, 6 Quellen
Aufsatz (Konferenz)
Englisch
A car detection system using neural network for image processing
Kraftfahrwesen | 1992
|Image processing using evolutive neural network
SPIE | 2006
|Image Processing using an Image Approximation Neural Network
British Library Conference Proceedings | 1994
|