1. 陕西科技大学 电气与信息工程学院,陕西,西安,710021
2. 西安微电子技术研究所,陕西,西安,710054
纸质出版:2019
移动端阅览
史永胜, 洪鑫扬, 段青亚, 马毅超, 陶亚凡, 何立风. 图像欧拉数计算的FPGA加速设计[J]. 微电子学与计算机, 2019, 36(8): 6-9.
SHI Yong-sheng, HONG Xin-yang, Duan Qing-ya, et al. A FPGA accelerator to compute image euler number[J]. 2019, 36(8): 6-9.
欧拉数是二值图像重要特征之一
为进一步提高现有二值图像欧拉数算法的效率
根据图论中的欧拉定理
在GRAY提出的基于四方块类型的欧拉数算法的基础上
提出一种欧拉数计算的FPGA加速设计.充分利用FPGA的并行处理能力
采用三级流水线的设计进行处理
使得计算欧拉数的时间几乎等于从RAM中读取一幅图的时间
极大提高了欧拉数计算的速度.最后使用41张随机二值噪声图像进行测试
结果表明
PC上的先进算法的处理时间随目标像素的密度的改变有很大的变化
最慢时间为最快时间的5倍左右
而本文设计的FPGA加速设计中
处理一幅图像的时间与目标像素无关仅与图片大小有关
且与理论上读取一幅图片的时间非常接近
速度比PC上先进的算法的平均速度快了20多倍
且成本功耗均远低于PC.
Euler number is an important feature of a binary image. Now
this paper proposes an efficient FPGA accelerator to improving the current computing efficiency which is designed for computing the Euler number based on the Euler Method of Gray System Model
according to the Euler theorem in the graph theory. Making good use of the parallel processing capability of FPGA
adopting design of three stage pipeline
the speed of FPGA accelerator computing the Euler number is improved greatly. The random 41 noisy binary images tests turn out that processing time of the advanced algorithm in PC is sensitive to the target pixels density change and the slowest is five times slower than the fastest. However
processing a image in FPGA accelerator is no concern with pixels density and low power depletion fast as read a image in theoretically
which is 20 times faster than advanced algorithm in PC.
JUANLDS, AZUELAJHS. On the computation of the Euler number of a binary object[J].Pattern Recognition, 1996, 29(3):471-476. DOI: 471
DEY S, BHATTACHARYA B B, KUNDU M K, et al.A fast algorithm for computing the Euler number of an image and its VLSI implementation[C]//Proc. the 13th Int. Conf. India: VLSI Design, 2000: 330-335.
SRIHARI S N.Document image understanding[C]// Proc.ACM.Dallas TX : IEEE Joint Fall Computer Conference, 1986: 87-95.
GONZALEZ R C, WOODS R E. Digital Image Processing (3rd Edition)[M]. Prentice-Hall, Inc. 2007.
BISHUNA, BHATTACHARYBB, KUNDUMK, et al. Apipeline architecture for computing the Euler number of a binary image[J].Journal of Systems Architecture, 2005, 51(8):470-487. DOI: 470
ROSIN P L, ELLIS T. Image difference threshold strategies and shadow detection[C]// Machine Vision Conference, British: BMVA Press Surrey, 1995: 347-356.
杨帆, 张皓, 马新文, 等.基于FPGA的图像处理系统[J].华中科技大学学报, 2015, 43(2):119-123. DOI: 119
雷国伟, 郭东辉.基于CNN二值图像处理的欧拉数计算[J]哈尔滨工程大学学报, 2006, 27(z1):202-204. DOI: 202
HASHIZUME A, SUZUKI R, YOKOUCHI H, et al.An Algorithm of Automated RBC Classification and Its Evaluation[J].Medical Electronics and Bio Engineering, 1990, 28(1): 25-32. DOI: 25
BISHNU A, BHATTACHARYA B B, KUNDU M K, et al. Euler vector for search and retrieval of gray-tone images[J]. IEEE Transactions on Systems Man % Cybernetics Part B Cybernetics A Publication of the IEEE Systems Man % Cybernetics Society, 2005, 35(4):801-12. DOI: 801
HE L F, CHAO Y Y, SUZUKI K.An Algorithm for Connected Component Labeling, Hole Labeling and Euler Number Computing[J]. Journal of Computer Science and Technology (English Language Edition), 2013, 28(3):468-478. DOI: 468
HE L, CHAO Y, SUZUKI K, et al.An efficient first-scan method for label-equivalence-based labeling algorithms[J]. Pattern Recogn. Lett. 2010, 31(1): 28-35. DOI: 28
姚斌, 何立风, 康世英, 等.一种快速二值图像欧拉数算法[J].微电子与计算机, 2017, 34(7):46-50. DOI: 46
姚斌, 何立风, 康世英, 等.基于图论的二值图像八邻接欧拉数算法[J].海军工程大学学报, 2016, 28(5):36-39. DOI: 36
GRAY S B. Local properties of binary images in two dimensions [J]. IEEE Transa.Computer, 1971, 5(C-20):551-561. DOI: 551
0
浏览量
10
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010602201714号