文章摘要
宋杨,殷焕松.基于FPGA的实时图像预处理系统研究[J].安徽建筑大学学报,2024,32(2):49-54,66
基于FPGA的实时图像预处理系统研究
Research on FPGA-Based Real-time Image Preprocessing System
  
DOI:
中文关键词: FPGA  双边滤波  边缘检测  视频图像处理  图像去噪
英文关键词: FPGA  bilateral filtering  edge detection  video image processing  image denoising
基金项目:高校优秀拔尖人才培育资助项目(gxyq2019056);安徽省住房城乡建设科学技术计划项目(2021-YF24、2022-YF168)
作者单位
宋杨 安徽建筑大学 电子与信息工程学院安徽 合肥 230601 
殷焕松 安徽建筑大学 电子与信息工程学院安徽 合肥 230601 
摘要点击次数: 80
全文下载次数: 0
中文摘要:
      考虑到对图像去噪的实时性要求和对图像边缘保护要求越来越高,充分利用 FPGA的并行数据处理能力强和逻辑资源丰富等特点,设计了一套基于FPGA的实时图像预处理系统,从而弥补PC处理速度慢的缺点。针对图像采集过程中出现的各种随机噪声,在硬件开发平台上基于FPGA设计了双边滤波算法、Sobel边缘检测算法、膨胀算法与腐蚀算法,并将处理后的图像显示在TFT LCD显示屏上。仿真及实验显示,处理640 × 480分辨率图像在PC上耗时 1.52 s,使用本文方案耗时 14.3 ms,表明该方案既可以降噪平衡,又保证了边缘效果,预处理速度达到了30帧/s。
英文摘要:
      Considering the increasing real-time requirements for image denoising and image edge protection, and the characteristics ofFPGA such as strong parallel data processing ability and abundant logic resources, a real-time image preprocessing system based onFPGA was designed, which could make up for the slow PC processing speed. To solve various random noises in the image acquisitionprocess, the bilateral filtering algorithm, Sobel edge detection algorithm, expansion algorithm and corrosion algorithm were combinedbased on FPGA on the hardware development platform, and the processed images were displayed on the TFT LCD display. Thesimulation and experimental results showed that it took 1.52 s to process the 640 × 480 resolution image on the PC using Sobel edgedetection algorithm and thresholding algorithm, while only 14.3 ms using the proposed scheme, which can not only reduce the noise,but also keep the edge clear, and the preprocessing speed reached 30 fps.
查看全文   查看/发表评论  下载PDF阅读器
关闭

分享按钮