论文部分内容阅读
为了解决小波变换在图像实时处理系统中的瓶颈问题,提出了一种硬件实现9/7整形小波高速变换的方法。该方法使用新的小波变换结构和小波变换基,采用一种基于行列同时变换的多级同时变换方式,即进行多级同时变换,且每一级的行列变换也同时进行。使用现场可编程逻辑门阵列(FPGA,Field Programmable Gate Array)的IP核实现片内缓存,降低了设计的复杂度,实现了图像的高速分解。整个设计采用VHDL对算法完成建模和实现,综合和仿真结果表明,该系统占用的资源少,分解速度很快,实现了高速的图像数据流输出,可应用到图像压缩的很多领域。
In order to solve the bottleneck of wavelet transform in image real-time processing system, a method of hardware realization of 9/7 shaping wavelet high-speed transform is proposed. The method uses a new wavelet transform structure and a wavelet transform basis, adopts a multi-level simultaneous transformation method based on the simultaneous transformation of ranks, that is, performs multiple simultaneous transformations, and the ranks transformations of each stage are also performed simultaneously. By using field programmable gate array (FPGA) Field-programmable Gate Array (FPGA) IP core to implement on-chip cache, the design complexity is reduced, and high-speed image decomposition is realized. The whole design uses VHDL to complete the modeling and implementation of the algorithm. The synthesis and simulation results show that the system occupies less resources and decomposes quickly, and achieves high-speed image data stream output, which can be applied to many fields of image compression.