基于OMAP3530的车道线识别算法研究与实现-电子与通信工程专业论文.docxVIP

中北大学学位论文 中北大学学位论文 The lane line recognition algorithm research and implementation based on OMAP3530 Abstract Vision navigation is one of the key technology of intelligent vehicle navigation, intelligent vehicles vision navigation system, and this paper studied the lane line recognition algorithm based on image detection. Based on OMAP3530 panel, the lane line recognition algorithm based on monocular vision were studied. Acquisition of road images using an external CMOS camera and image processing technology to complete the lane line detection and extraction, provide the basis for visual navigation decision further. Based on the lane line identification system real-time performance, reliability and stability requirement, this paper studies the OMAP3530 framework on the ARM embedded Linux system design and video image acquisition subsystem, DSP side lane line recognition algorithm and the communication mechanism of ARM and DSP kernel DSP/BIOS Link and so on, puts forward the solution based on OMAP3530 lane line identification. Series of OMAP ARM microprocessor is mainly studied the embedded Linux system of each module, including MLO, uboot, kernel, file system, and the USB video device driver module, and based on V4L2 video image acquisition module, and transplanted them into OMAP3530 platform, improve the real-time performance of the system and the ability to respond. The design of the video image acquisition module can be real-time acquisition of BMP and YUV format images. Studies the lane line identification algorithm in the vision navigation. Main process is: the image gray, the image filtering, image edge enhancement, image edge detection, lane line extraction, finally obtained the lane line clearly. Then the identification algorithm is applied to the structured road, obtained the expected effect. Experimental results show that the algorithm can ensure the accuracy of the lane line recognition, real-time and robustness, can satisfy the actual needs. Keywords：Visi