基于FPGA的多用途提升小波变换核

根据小波变换的框架式结构,提出了一种在FPGA上完全依靠重构来实现不同提升小波变换核的方法。根据变换特性和重构要求,采用了由下至上的结构体设计方法,将每个提升步骤用可编程的参数来表示,以保证每个提升步骤均可重构。给出了提升小波变换核在FPGA上的结构体,在逻辑综合时按不同小波的要求,改变参数可得到不同的结果。仿真实验表明,提升小波变换核的结构体在处理能力和功耗方面都取得了很好的效果,尤其在处理速度上,能满足图像处理实时性要求。

Multi-purpose lifting wavelet transform kernels based on FPGA

According to the framework of wavelet transform, a fully reconfigurable multi-purpose lifting kernel on FPGA was presented. From transform characteristics and reconfiguration requirement, with a bottom-up approach in design architecture, every lifting step can be represented by programmable parameters, which can guarantee lifting scheme reconfigurability. In addition, the architecture of lifting wavelet transform kernel was given. Different result was obtained through adjusting parameters in logical synthesis. Simulations illustrated that this lifting wavelet transform architecture possesses strong processing ability and low power consumption in real time image processing.