基于FPGA的图像边缘保护高斯滤波算法实现  

Realization of image edge protection Gauss filter algorithm based on FPGA

作　　者：康宇 赵冬青[1] 上官鹏 储成群[1] KANG Yu;ZHAO Dongqing;SHANGGUAN Peng;CHU Chengqun(State Key Laboratory of Electronic Testing Technology,North University of China,Taiyuan 030051,China;Beijing Institute of Control and Electronic Technology,Beijing 100032,China)

机构地区：[1]中北大学电子测试技术国家重点实验室,山西太原030051 [2]北京控制与电子技术研究所,北京100032

出　　处：《电子设计工程》2021年第6期94-98,110,共6页Electronic Design Engineering

摘　　要：针对传统高斯滤波在图像边缘细节保留上的不足,实现了一种基于FPGA的改进高斯滤波算法,在保留图像边缘信息的同时滤除高频噪声,并实现高速实时图像处理。通过对当前图像像素点与由图像全局求取的梯度阈值的比对,以确定其为噪点或边缘点,并做区分处理。边缘点将不予以处理,而噪点则通过权重模板进行滤波处理。经仿真与测试,该算法在FPGA中实现占用资源少,运行速度快,可实现图像高速实时处理,图像输出延时仅一行像素时间,经过多组图像分析比较,相比当前基于FPGA的高斯滤波,低噪环境下图像PSNR可提高逾6%,MSE可提高30%以上。Aiming at the shortcomings of traditional Gauss filter in image edge detail preservation,an improved Gauss filter algorithm is implemented based on FPGA,which can remove high⁃frequency noise while preserving image edge information and realize high⁃speed real⁃time image processing.One pixel can be determined as the noise or edge point through the comparison between the current image pixel and the gradient threshold obtained by the image global and do the distinction processing.The edge points will not be processed,while the noise will be filtered through the weight template.After simulation and test,the algorithm can achieve high⁃speed real⁃time image processing with low resource consumption and have fast running speed in FPGA.The image output delay is only one row of pixel time.After analysis and comparison of multiple groups of images,compared with the current Gaussian filter based on FPGA,the PSNR and MSE of image in low noise environment can be increased by more than 6%and 30%respectively.

关 键 词：高斯滤波 实时图像处理 图像边缘 噪点检测 

分 类 号：TN919[电子电信—通信与信息系统]