针对串行架构的处理器如 ARM和 DSP等实现双边滤波算法消耗资源过多的情况, 提出了一种基于 FPGA加速的双边滤波去噪的实现方法。根据双边滤波算法的原理以及 FPGA的硬件资源, 充分发挥 FPGA在并行计算上的优势, 在滤波器的系数生成和除法操作上进行了优化。在减少了 FPGA所消耗的资源的同时, 保证了图像去除噪声的实时性, 从而为后期的视频采集与传输系统奠定了一个良好的开端。

Considering that general serial processors, such as ARM and DSP, consume too many resources, we propose a method for a bilateral filter based on FPGA acceleration. According to the theory of a bilateral filter and the hardware resources that an FPGA can provide, we optimize the coefficient generation and division operation of the filter in the FPGA, which takes full advantage of the parallel computing capabilities in the FPGA. This ensures that the image is filtered in real-time, while reducing the resources consumed by the FPGA, and provides a good foundation for the later design of the video capture and transmission systems.