【摘要】 边缘信息是数字图像信息的重要组成部分,传统的边缘检测程序会因串行执行指令而导致检测效率下降。为了提高图像边缘信息的处理能力,提出一种基于现场可编程门阵列的改进Sobel算子的边缘检测方法,其中,Sobel算法的卷积核从2方向扩展到8方向,并具备阈值自适应能力。板级测试结果表明,基于现场可编程门阵列的改进Sobel算法边缘检测方法充分发挥FPGA并行处理的优势,能够高效准确地检测图像的边缘信息,图像层次感强、成像稳定,在图像的并行处理中有明显的优势。更多还原

【Abstract】 Edge information is an important part of digital image information. The traditional edge detection program reduces the detection efficiency because of its serial execution instructions. In order to improve the processing ability of image edge information, this paper proposed an improved Sobel operator based on FPGA(Field Programmable Gate Arrays). The convolution kernel of Sobel algorithm was extended from 2 directions to 8 directions and it had threshold adaptive capability. Board-level experiments show that the FPGA-based improved Sobel algorithm edge detection method gives full play to the advantages of FPGA parallel processing and it can efficiently and accurately detect the edge information of images. It not only has a strong sense of image hierarchy and stable imaging, but also has obvious advantages in parallel processing of images.更多还原