A Vehicle Detection Technique for Traffic Management using Image Processing

Traffic congestion is an unfortunate trusted companion of every urban dweller living in busy non-static cities. In order to manage traffic both manually and automatically, accurate vehicle detection is very important. In this paper, a model has been proposed for vehicle detection to control traffic using image processing. At first, this system converts RGB road images to HSV images. Then, it analyses the value readings which identify the brightness of the images, to determine whether it is a day time or night time image by comparing the value readings with a calculated threshold parameter. At this stage, two different methodologies have been used to detect day time and night time vehicles. During day time, comparison is done between foreground image with the background image to extract the vehicles. Then object counting methodology is applied to count the number of vehicles. On the other hand, during night time the intensity of the image is analyzed to differentiate between headlights and ambient light. Finally, another object counting methodology is used to count the number of vehicles. This proposed model is tested with different dataset and it exhibits average 95% accuracy for day and night time.