Vision-Based Vehicle Detection based on Embedded Systems

Over the last decade have seen growing importance placed on research in vehicle detection. While implementing the vehicle detecting system, the space of the car and the cost of the hardware platform are the decisive consideration. In this paper, we proposed a modified vehicle detection algorithm to make it can be implemented on the embedded platform. The embedded platform is based on an INTEL XScale PXA270 SoC and limited hardware resources. An application program realizes the proposed algorithm and integrates V4L2 and MiniGUI open source codes under LINUX operating system. The proposed algorithm verifies the preceding vehicles driving on the highway at daytime. In our experiments, the system can correctly verify and track vehicles. And the processing time of this system is less than the human reaction time. Keywords— Vision-based Vehicle Detection, Embedded System.