Abstract
The large increase in traffic density during the last 25 years in the US, Europe, and Japan has required the automation of some functions of traffic control and monitoring. Buried magnetic loops are widely used for this purpose, but there are functions such as incident detection, vehicle tracking, etc., which are not implementable with them. In addition, magnetic loops are not flexible. Image processing represents a potentially much more powerful tool for this application. Its main disadvantage is the high initial cost. Research in the area of machine vision for traffic monitoring and control is being actively pursued in several countries. This paper surveys current efforts.

This publication has 3 references indexed in Scilit: