| Lane Departure Warning System(LDWS), as an important active safety technology on vehicle, plays a vital role in reducing traffic accidents caused by lane departure. In terms of preventing lane departure domestic and foreign have carried out considerable and helpful research and made a breakthrough in sensor technology, radar, laser, ultrasound, infrared, machine vision and so on. In this paper, the front-end of road image processing system of lane departure warning is studied based on DM642 hardware platform, and its effectiveness is verified by testing.In this paper, the hardware platform is TMS320DM642 experimental board developed by ShenZhen QiXiangDa to achieve road image acquisition and image pre-processing display, equipped with other hardware such as front-end road image pickup devices, monitor and so on. After building the hardware platform and Integrated Development Environment CCS, the program is solidified in foreign expansion Flash to achieve successfully offline operation of system according to DM642’s own characteristics.In this paper, the front-end image processing technology of the system is studied and the road image pickup, decoding principle and transmission between DM642 and video image are analyzed. First of all, the format of the system video acquisition is set, and the acquisition process for the CCD camera is drew an analogy. The function of I2 C bus of the system as well as the design and implementation of bus switching and time division multiplex access is described. Following, video decoding module circuit of the system which contains TVP5150 video input circuit, the configuration and connection of DM642 and the decoding chip, connection of data video streaming between DM642 and TVP5150 is analyzed and designed. The relevant registers of TVP5150 which are involved in decoding process are configured and their function is listed. Based on this, the video decoding process is depth analyzed from image digitization, conversion of RGB and YUV, the sampling of standard BT.656 YUV4:2:2 digital video streaming and transmission format etc. Finally, how DM642 video port achieves and captures YUV4:2:2 digital video streaming, how to realize the move of data from FIFO buffer of video ports to DM642 memory area through EDMA interrupt are studied.By the real vehicle experiment, some image preprocessing algorithm, such as gray scale, binarization of road image etc have been achieved and the reliability of hardware platform and software algorithm are verified. |
