Implementation and Optimization of Canny Edge Detection Algorithm on FT Platform

doi:10.19678/j.issn.1000-3428.0059943

Computer Engineering ›› 2021, Vol. 47 ›› Issue (7): 37-43. doi: 10.19678/j.issn.1000-3428.0059943

• Research Hotspots and Reviews • Previous Articles Next Articles

Implementation and Optimization of Canny Edge Detection Algorithm on FT Platform

GUO Hengliang¹, CHAI Xiaonan², HAN Lin¹, HE Xiaohui³, SHANG Jiandong¹

1. Henan Province Supercomputing Center, Zhengzhou University, Zhengzhou 450000, China;
2. School of Information Engineering, Zhengzhou University, Zhengzhou 450000, China;
3. School of Earth Science and Technology, Zhengzhou University, Zhengzhou 450000, China

Received:2020-11-09 Revised:2020-12-10 Published:2020-12-15

Canny边缘检测算法在飞腾平台上的实现与优化

郭恒亮¹, 柴晓楠², 韩林¹, 赫晓慧³, 商建东¹

1. 郑州大学河南省超级计算中心, 郑州 450000;
2. 郑州大学信息工程学院, 郑州 450000;
3. 郑州大学地球科学与技术学院, 郑州 450000

作者简介:郭恒亮(1971-),男,副教授,主研方向为智慧城市、地理信息系统、高性能计算;柴晓楠,硕士研究生;韩林(通信作者),副教授;赫晓慧、商建东,教授。
基金资助:
国家重点研发计划（2018YFB0505000）。

Abstract

Abstract: In order to support the underlying image library on the FT DSP platform,and reduce the time consumed by the calculation in the Canny edge detection algorithm,an algorithm for parallel Canny gradient computing based on FT-M7002 is proposed.On the basis of FT-M7002 high-performance processing architecture,Single Instruction Multiple Data(SIMD) is vectorized to enhance the parallel processing of the instructions of DSP cores.According to the hierarchical structure features of the vector memory of FT-M7002,the mode of data memory access of the Canny parallel gradient computing algorithm is analyzed.The first address offset is used to deal with discontinuous data memory access,and data transmission and data calculation is completed by means of double buffering mode.Experimental results show that when reaching the same detection accuracy as the original Canny algorithm,the proposed algorithm improves the overall running speed by 1.490~2.112 times when the size of convolution core is 3×3,5×5,and 7×7,bridging the performance gap with the mainstream accelerators in digital image processing.

Key words: FT-M7002 processor, Canny edge detection, parallel gradient computing, memory access optimization, double buffering mode

摘要： 为实现国产飞腾DSP平台对底层图像库的支持，针对原始Canny边缘检测算法计算时间过长的问题，设计一种面向FT-M7002平台的Canny梯度计算并行算法。基于FT-M7002高性能处理架构，采用单指令流多数据流向量化方式增强DSP内核指令的并行处理能力，根据FT-M7002平台向量存储器的层次结构特征，分析Canny梯度计算并行算法的访存模式，通过首地址偏移取址解决不连续访存问题，并结合双缓冲方式完成数据传输与数据计算。实验结果表明，在与原始Canny算法具有相同检测精度的情况下，该算法在卷积核大小为3×3、5×5、7×7时整体运行速度提升了1.490~2.112倍，缩小了与主流加速器件在数字图像处理领域的性能差距。

关键词: FT-M7002处理器, Canny边缘检测, 梯度计算并行, 访存优化, 双缓冲方式

CLC Number:

TP391

GUO Hengliang, CHAI Xiaonan, HAN Lin, HE Xiaohui, SHANG Jiandong. Implementation and Optimization of Canny Edge Detection Algorithm on FT Platform[J]. Computer Engineering, 2021, 47(7): 37-43.

郭恒亮, 柴晓楠, 韩林, 赫晓慧, 商建东. Canny边缘检测算法在飞腾平台上的实现与优化[J]. 计算机工程, 2021, 47(7): 37-43.

/ / Recommend / Download Citations

URL: http://www.ecice06.com/EN/10.19678/j.issn.1000-3428.0059943

http://www.ecice06.com/EN/Y2021/V47/I7/37

Figures/Tables 18

References

[1] 李飞,宾洋,罗文广,等.基于TMSDM6437的车道线检测[J].广西科技大学学报,2018,29(1):43-48. LI F,BIN Y,LUO W G,et al.Lane line detection based on TMSDM6437[J].Journal of Guangxi University of Science and Technology,2018,29(1):43-48.(in Chinese)
[2] 卢俊,张保明,黄薇,等.基于GPU的遥感影像数据融合IHS变换算法[J].计算机工程,2009,35(7):261-263. LU J,ZHANG B M,HUANG W,et al.IHS transform algorithm of remote sensing image data fusion based on GPU[J].Computer Engineering,2009,35(7):261-263. (in Chinese)
[3] LIU F,GUAN G X,ZHAO H M,et al.Fast processing of UAV remote sensing image based on DSP hardware simulation platform[EB/OL].[2020-10-04].https://www.researchgate.net/publication/340737558_Fast_Processing_of_UAV_Remote_Sensing_Image_Based_on_DSP_Hardware_Simulation_Platform.
[4] OKAMOTO T,KOIDE T,YOSHIDA S,et al.Implementation of computer-aided diagnosis system on customizable DSP core for colorectal endoscopic images with CNN features and SVM[C]//Proceedings of the 28th International Conference on Microelectronics.Washington D.C.,USA:IEEE Press,2019:1663-1666.
[5] AGGARWAL V,GUPTA A.Integrating morphological edge detection and mutual information for nonrigid registration of medical images[J].Current Medical Imaging Reviews,2019,15(3):292-300.
[6] ZHOU X L,XU L M,WANG J.Road crack edge detection based on wavelet transform[EB/OL].[2020-10-04].https://iopscience.iop.org/article/10.1088/1755-1315/237/3/032132/pdf.
[7] CUI J C,TIAN K.Edge detection algorithm optimization and simulation based on machine learning method and image depth information[J].IEEE Sensors Journal,2020,20(20):11770-11777.
[8] 宋琦.DSP芯片在实时图像处理系统中的应用分析[J].电子技术与软件工程,2015(9):86. SONG Q.Application analysis of DSP chip in real-time image processing system[J].Electronic Technology and Software Engineering,2015(9):86.(in Chinese)
[9] PR Newswire.CEVA announces industry's first high performance sensor hub DSP architecture[EB/OL].[2020-10-04].https://www.baidu.com/link?url=GLlzb7iS0JKumoTpnYQW370As9qsRZg90UHlhRTSN1-wBhtmNK9-L1dISRS8qFymXV8ZgoetG4dlDDHfuLohrq&wd=&eqid=db5b583e000301630000000460a759c3.
[10] MODY M,HARIYANI H,BALAGOPALAKRISHNAN A,et al.GPU assist using DSP pre-processor[C]//Proceedings of 2020 IEEE International Conference on Electronics,Computing and Communication Technologies.Washington D.C.,USA:IEEE Press,2020:1-4.
[11] NAJOUI M,BAHTAT M,HATIM A.VLIW DSP-based low-level instruction scheme of givens QR decomposition for real-time processing[J].Journal of Circuits Systems and Computers,2017,26(9):478-504.
[12] KUMAR A,RAHEJA S.Edge detection using guided image filtering and enhanced ant colony optimization[J].Procedia Computer Science,2020,173:8-17.
[13] SENGUPTA S,MITTAL N,MODI M.Improved skin lesion edge detection method using ant colony optimization[J].Skin Research and Technology,2019,25(6):846-856.
[14] DAGAR N S,DAHIYA P K.Edge detection technique using binary particle swarm optimization[J].Procedia Computer Science,2020,167:1421-1436.
[15] 陈超.多阈值优化的运动图像轮廓特征提取方法[J].沈阳工业大学学报,2019,41(3):315-319. CHEN C.Extraction method of contour features by multi-threshold optimization for motion images[J].Journal of Shenyang University of Technology,2019,41(3):315-319.(in Chinese)
[16] WANG S Q,LIANG S,PENG F,et al.Image edge detection algorithm based on fuzzy set[J].Journal of Intelligent & Fuzzy Systems,2020,38(4):3557-3566.
[17] HOU Z R.Dual threshold and edge image optimization of Canny algorithm in the road extraction from remote sensing image[J].International Journal of Earth Sciences and Engineering,2015,8(1):188-194.
[18] WU G H,YANG D Y,CHANG C,et al.Optimizations of Canny edge detection in ghost imaging[J].Journal of the Korean Physical Society,2019,75(3):223-228.
[19] SHI Q N,AN J C,GAGNON K K,et al.Image edge detection based on the Canny edge and the ant colony optimization algorithm[C]//Proceedings of the 12th International Congress on Image and Signal Processing.Washington D.C.,USA:IEEE Press,2019:12-18.
[20] 孙广辉,扈啸,王蕊.基于FT-M7002的OpenCV移植与优化[C]//第二十二届计算机工程与工艺年会暨第八届微处理器技术论坛论文集.长沙:中南大学出版社,2018:113-121. SUN G H,HU X,WANG R.OpenCV transplantation and optimization based on FT-M7002[C]//Proceedings of the 22nd Annual Conference of Computer Engineering and Technology and the 8th Microprocessor Technology Forum.Changsha:Central South University Press,2018:113-121.(in Chinese)
[21] 宋贵环.YHFT-Matrix2编译器SIMD优化技术研究与实现[D].长沙:国防科技大学,2014. SONG G H.Research and realization of YHFT-Matrix2 compiler SIMD Optimization technology[D].Changsha:National University of Defense Technology,2014.(in Chinese)

Please choose a citation manager

Content to export