基于FPGA和DSP的车牌识别系统的研究

Study of FPGA & DSP-based vehicle license plate recognition system

摘要:

针对常见的车牌识别系统依赖上位机成本高,并且体积较大不利于野外工作的问题,采用了一种以DSP和FPGA为开发平台的车牌识别系统的设计方案;利用FPGA控制CMOS图像传感器进行车牌图像数据的采集,经过图像格式转换之后传送到DSP中,进行车牌区域定位、字符分割、车牌识别等操作,最后将识别结果传送到LCD中显示;系统采用CCS(Code Composer Studio)集成开发环境,利用VC++语言编程实现车牌识别算法;通过嵌入式机器视觉库EMCV实现识别算法在DSP中的移植,并在LCD上创建工作窗口显示识别结果,检测速度快,准确率高;实验结果表明,该车牌识别系统的设计方案识别结果准确,功能稳定可靠,有效地解决了常见车牌识别系统成本高,体积大不利于野外工作的问题,达到了设计目的。

Abstract:

The common license plate recognition system should work with the PC, costing much. What’s more, it occupy large space, inconvenient to use outdoors. To solve these problems, this paper introduces a design of the license plate recognition system based on DSP and FPGA as the core. The FPGA control system is used to control the CMOS image sensor acquiring license plate image data, and data transmission to DSP of the format-conversed image. Then the DSP control the operation of the license plate location, character segmentation, license plate recognition, and display of the identification results in the LCD. The system uses CCS(Composer Studio Code) integrated development environment, using VC++ language programming to achieve the license plate recognition algorithm. The algorithm transplantation into the DSP is realized by the embedded machine vision library, and the results are displayed in the work window created in the LCD, with the advantage of fast detection speed and high correction rate. We can conclude from the experimental results that the design scheme of the license plate recognition system is accurate and reliable. And it successfully handled the problems listed in the beginning, achieve our design purpose.