基于FPGA实现二维卷积滤波器的图像处理

Realization of FPGA⁃based image processing of 2D convolution filter

目前的二维卷积滤波图像处理方法本质上都是基于PC端进行算法仿真,而在将仿真算法映射到硬件平台进行实现时,处理效果不佳。针对这一问题,文中提出了基于FPGA构建多种二维卷积滤波器共同实现方案。所提方案是基于FPGA平台,通过在FPGA上进行Shift⁃RAM搭建,基于二维卷积核算法,做到同时实现多种图像处理,比如RGB彩色图像处理,灰度处理和图像边缘检测,并在硬件设计上对图像边界处理的映射算法进行完善。实验结果表明,所提案能够在FPGA平台上很好地实现多种基于二维卷积的仿真算法,同时,在保证处理效果的前提下,提高了处理效率。

As the current image processing method of 2D convolution filtering is essentially based on the PC⁃side to make algorithm simulation,the processing effect is not good when the simulation algorithm is mapped to the hardware platform.On this basis,a common implementation scheme of constructing multiple types of 2D convolution filters on the basis of FPGA is proposed.The new scheme is based on the FPGA platform.By establishing the Shift⁃RAM on the FPGA,the multiple types of image processing ways are realized on the basis of 2D convolution kernel algorithm,such as RGB color image processing,gray processing and image edge detection.The mapping algorithm for image boundary processing is improved in the hardware design.The experimental results show that the proposed scheme can achieve a variety of simulation algorithms based on 2D convolution on the FPGA platform,and improve the processing efficiency while ensuring the treatment effect.