Multithreading Application for Counting Vehicle by Using Background Subtraction Method

Abstract

Image and video processing has become important part in intelligent transportation system (ITS) application, especially for collecting road traffic data. Pictures that already collected by a charged coupled device (CCD) camera usually being processed by several image processing algorithms and the application’s code will be executed in a large number of iteration because many algorithms are getting involved in processing the frame which captured by the camera. Typical application will process the first frame until finish and then continue to the next frame, so the application must wait until the first frame being processed. If the algorithms that executed quite complex and have a significant running time there will be a dropped frame and the time difference between data acquisition and real time video is divided by large margin. We proposed an implementation of multithreading to boost the application performance so the data can be acquire in real time and every new frame could be processed in short time. The application performance before and after using a multithreading is known by comparing the data acquisition time that stored in the database. The application effectiveness could define by running a multiple video streaming in same resolution.