基于FPGA的5/3提升小波优化算法

为了实现线阵CCD空间相机图像的实时压缩处理,在提升算法的基础上,提出了一种适用于FPGA的二维提升小波变换结构与实现方案.该系统利用FPGA片内的存储资源,采用乒乓操作实现了行列变换之间的数据缓存传输,降低了功耗,提高了硬件利用率和运算速度.并且为了适应硬件实现速度,在进行小波边界处理时不需要额外的边界延拓过程,很大程度上降低了算法的复杂度;整个模块采用verilog HDL语言进行设计,并在QuestaSim下进行了仿真试验.实验结果表明,该系统工作稳定可靠,完全满足实时处理的要求,并适用于JPEG2000的多级二维5/3小波变换.

The Optimization Algorithm of 5/3 Lifting Wavelet Based on FPGA

In order to implement the real-time image data compression processing for line CCD space camera, the paper proposed a new method for lifting wavelet transform, which is based on the algorithm structure of wavelet transform. By using the abundant on-chip memories and ping-pong scheme, it can implement the data transmission between row transform and column transform. The method can reduce the power and enhance the speed of computation. It abandons the symmetric extension on the boundary, which can simply the complexity of computation. The module is designed by verilog HDL, simulated by Questasim. The results indicate that the system can work steadily, and it is reasonable to implement real-time processing and multilevel 2-D 5/3 wavelet transform.