Vehicle classification and counting for vehicle census

Vehicle classification has been significantly important to vehicle census as it provides traffic count information to reflect the traffic density of a particular roadway. however,it has been a time consuming and sophisticated task to classify different vehicles into the desired category. besides,the hardware-based technique used for classification leads to high cost of implementation and maintenance. thus,we proposed an image processing based solution to extract the features of each vehicle in the traffic scene. the proposed framework incorporates a combination of detection,tracking and classification of vehicle to ensure high accuracy and performance for vehicle census. experimental results show that our proposed framework can be applicable in real world applications.