DSP车牌识别系统.docx
版权申诉
4 浏览量
2022-06-08
11:22:52
上传
评论
收藏 836KB DOCX 举报
摘要
本文介绍了一个以 TI 公司的 TMS320DM642 为核心芯片的 DSP 车牌识别系统设计、实现
和优化。该系统首先通过摄像机拍摄车辆的视频，输入视频信号到DSP 板卡;然后从输入的视
频信号中捕捉图像，识别图像中的车牌的类型、颜色和号码，最后通过串口把识别结果传到
PC 机。
该系统主要分为三个部分:摄像机、DSP 系统和 PC 端软件。文章主要介绍了 DSP 系统中
DSP 车牌识别软件的实现和优化。由于 DSP 系统与普通 PC 机的不同，文章中详细说明了为
了提高运行的速度，对 DSP 车牌识别软件进行的各种优化。这些优化主要包括，提高并行性、
减少运算和使用 TI 提供的经过优化的库等等。经过这些优化，使DSP 车牌识别系统能在（不
到 0.4 秒）很短时间内完成一个车牌的识别及其他处理，满足了实际应用的要求。
关键字:车牌识别，DSP 芯片，嵌入式系统，优化
ABSTRACT
This paper introduces the designation, implementation and optimization of a LPR System based
on the TI's DSP chip TMS320DM642. First, this system uses video camera to record video. Then the
DSP board, which uses the video signals as input, captures image from the video, recognizes the
license plate' s type, color and number, and uses JPEG algorithm to compress the image. At last, the
result is transmitted to PC by serial port。
This system is divided into three parts: video camera, DSP system, PC-end software. The paper
mainly explains the optimization of LPR software running on DSP board. Because of the difference
between DSP system and PC, we have to use many ways to optimize the software. We improved the
parallelism of code, reduced the operations and used the optimized libraries supplied by TI. After the
optimization, the system can work faster than before and it achieves request for practical application.
Keywords: LPR, DSP Chip, Embedded System, Optimization
资源评论
