基于SOPC的智能车辆道路识别与跟踪

Road recognition and tracking for intelligent vehicle based on SOPC

对基于PC的道路识别系统进行改进,利用SOPC技术在以单片FPGA芯片为核心的硬件平台上,完成了包括3×3窗口模块、改进的快速中值滤波、基于自适应阈值的Sobel边缘检测和Hough变换等图像预处理功能。通过μC/OS-II嵌入式实时操作系统来调度系统初始化、数据接收、道路识别、车辆控制和状态监控等任务,保证了视频导航的实时性要求。静态道路图像识别和实际道路行驶测试表明,该系统在一定干扰的情况下,成功实现了视频导航功能,具有较强的道路识别跟踪能力,且在速度、价格、灵活性上有很大的优势。

The road recognition system based on PC machine is improved. On the hardware platform with single FPGA chip as the core and based on SOPC technology, image processing function modules have been accomplished, which include a 3 × 3 window, improved fast median filtering, Sobel edge detection based on self-adapting threshold, and Hough transform. Embedded real-time operating system μC/OS-II is used to schedule all the tasks, such as sys- tem initiation, data receiving, road recognition, vehicle control and state monitoring. Static state road image recogni- tion and practical road running test show that under certain disturbance, the system has achieved video navigation function successfully, has strong road recognition and tracking ability, and has great advantages on speed, price and flexibility.