FPGA与DSP在小波变换应用中的对比研究

Contrast Research on FPGA and DSP in the Application of Wavelet Transform

离散小波变换是一种典型的实时图像处理算法,近年来,已获得越来越多地关注。文章利用基于提升方案的二维离散小波变换思想,编写VHDL代码,在FPGA上实现;并基于这种思想编写C代码,在DSP上运行。最后对两种平台作出对比,以了解实时图像处理在两种不同平台上的效率。实验结果表明,FPGA在执行时间、功耗和硬件占用率方面优于传统的DSP,但DSP的开发难度低于FPGA。

The Discrete Wavelet Transform is a classic real-time imaging algorithm that is drawing the attention of engineers in recent years.In this paper,VHDL code is written in the FPGA using 2-D lifting based wavelet transform.And C code is written in DSP based on the same idea.Effects of real-time image processing on these two platforms are compared.Results of the experiment prove FPGA is super to traditional DSP processors in time execution,power dissipation,and hardware utilization.But DSP design is simpler than FPGA.