Automatic detection and counting of vehicles in an unsupervised videos on highways is very challenging in computer vision with important practical applications like monitoring traffic activities.Counting the number of objects is integral part of image processing.Knowing the number of objects present in the image is very useful in wide range of applications.The main objective of our study is to develop methodology for automatic vehicle detection and its counting on highways.A system has been developed to detect and count objects efficiently.We present a system for detecting and tracking vehicles in surveillance video which uses segmentation with initial background substraction using morphological operator to determine salient regions in sequence of video frames.