Abstract: Traffic intensities in Sri Lankan roads are ever increasing making current traffic light system ineffective. Then it is imperative to make new methods to control increasing traffic. This end could be achieved only by eliminating inherent defects in traffic light system. Among approaches suggested to address the problem of managing higher traffic intensities includes extract information about the traffic concentration by processing digital images. This study attempts to estimate the traffic intensity by using image processing which also a part of intelligent traffic control system adopted in the future. This study focused on eliminating the problem encountered in commercial image processing system which does not work well in traffic congestions. The difficulty arrives when vehicles are exposed partially and appears differently under different light and environment conditions. This study proposed to develop an algorithm to detect vehicles from a stream of images extracted from a video. This algorithm will separate vehicles from their backgrounds and detect vehicles on the basis of morphological characteristics.