Vehicle Detection and Classification using Image processing

摘要

The number of vehicles has increased tremendously over the past decade. There are over 1 billion active vehicles all over the world and 60 to 70 million vehicles in India. Managing such traffic moments, providing sufficient parking lots is not an easy task. Vehicle counting and classification on busy streets will help the authorities to obtain traffic flow statistics and help them to understand and study the traffic patterns so that the can manage traffic in the most efficient way. The paper presents a way to detect, count and classify vehicles using image processing techniques. Although there has been a significant amount of research related to this, there is always a scope of improvement. The task of vehicle detection and counting is broken down into six steps: 1) Image Acquisition, 2) Image Analysis, 3) Object detection, 4) Counting, 5) Classification, 6) Display result. The algorithms which will be used to perform these tasks will includes vehicle detection and counting algorithm and road marking detection algorithm. This can also be used to monitor high ways, detect accidents, unrighteous stoppage of vehicles on roads, the traffic rules violators. Classification of vehicles will be done in one of the following categories: a) Bicycles and motorcycles, b) motor cars, c) minibus and pickup vans, d) buses trailers, trucks. This data will help to figure out the priority and maximum users of a road and design traffic patterns that will be beneficial to maximum.