快速鲁棒特征算法的CUDA加速优化

计算机科学 ›› 2014, Vol. 41 ›› Issue (4): 24-27.

快速鲁棒特征算法的CUDA加速优化

刘金硕,曾秋梅,邹斌,江庄毅,邓娟

武汉大学计算机学院武汉430072;武汉大学计算机学院武汉430072;武汉大学计算机学院武汉430072;武汉大学计算机学院武汉430072;武汉大学国际软件学院武汉430072

出版日期:2018-11-14 发布日期:2018-11-14
基金资助:
本文受国家自然科学基金(61303214),地震行业重大专项(201008007)资助

Speed-up Robust Feature Image Registration Algorithm Based on CUDA

LIU Jin-shuo,ZENG Qiu-mei,ZOU Bin,JIANG Zhuang-yi and DENG Juan

Online:2018-11-14 Published:2018-11-14

摘要/Abstract

摘要： 提出一种基于统一计算设备架构(Compute Unified Device Architecture,CUDA)的快速鲁棒特征(Speed-up Robust Feature,SURF)图像匹配算法。分析了SURF算法的并行性,在图像处理单元(Graphics Processing Unit,GPU)的线程映射和内存模型方面对算法的构建尺度空间、特征点提取、特征点主方向的确定、特征描述子的生成及特征匹配5个步骤进行CUDA加速优化。实验表明,相比适用于CPU的SURF算法,文中提出的适用于GPU的SURF算法在处理30MB的图片时性能提高了33倍。适用于GPU的SURF算法拓展了SURF算法在遥感等领域的快速应用,尤其是大影像的快速配准。

关键词: 快速鲁棒特征,CUDA,特征提取,影像匹配

Abstract: This paper proposed a speed-up robust feature image registration algorithm based on Compute Unified Device Architecture (CUDA)．We analyzed the parallelism of SURF algorithm,and optimized it by CUDA from thread mapping and memory model of the Graphics Processing Unit (GPU),aiming at the five steps of SURF algorithm including building scale space,extracting feature points,determining key direction of feature points,generating vector descriptor and feature matching．The experimental results show that the GPU implementation achieves 33times faster than the CPU implementation while processing an image of 30MB．The GPU implementation extends the application of SURF algorithm in fast processing of remote sensing image,especially in the quick image registration.

Key words: Speed-up robust feature,CUDA,Feature extraction,Image matching

刘金硕,曾秋梅,邹斌,江庄毅,邓娟. 快速鲁棒特征算法的CUDA加速优化[J]. 计算机科学, 2014, 41(4): 24-27. https://doi.org/

LIU Jin-shuo,ZENG Qiu-mei,ZOU Bin,JIANG Zhuang-yi and DENG Juan. Speed-up Robust Feature Image Registration Algorithm Based on CUDA[J]. Computer Science, 2014, 41(4): 24-27. https://doi.org/

参考文献

[1] Bay H,Tuytelaars T,Gool V L．SURF:Speeded Up RobustFeatures[C]∥Proc of European Conference on Computer Vision 2006．Austria:Graz,2006:404-417
[2] Bay H,Ess A,Gool V L．Speeded Up Robust Features (SURF)[J]．Computer Vision and Image Understanding,2008,0(3):346-359
[3] Sinha N S,Frahm J,Pollefeys M,et al.GPU-based Video Feature Tracking and Matching[R]．EDGE Workshop on Edge Computing Using New Commodity Architectures,2006
[4] Sinha N S,Frahm J,Pollefeys M,et al.Feature tracking andmatching in video using programmable graphics hardware[J]．Machine Vision and Applications,2011,2(1):207-217
[5] 林晓帆,林立文,邓涛,基于SURF描述子的遥感影像配准[J]．计算机工程,2010,36(12):216-218
[6] Murillo A C,Guerrero J J,Sagues C．Surf features for efficient robot localization with omnidirectional images [C]∥IEEE International Conference on Roboties and Automation．Roma,Italy,2007:3901-3907
[7] Arturo G,Oscar M,Monica B,et al．A comparative evaluation of interest point detectors and local descriptors for visual SLAM [J]．Machine Vision and Application,2010,21(6):905-920
[8] Valgren C,Lilienthal A．SIFT SURF and Seasons:Longtermoutdoor Localization Using Local Features[C]∥Proc of the 3^rd European Conf on Mobile Robots．Germany,2007:253-260
[9] 李慧,蔺启忠,刘庆杰．基于FAST和SURF的遥感影像自动配准方法[J]．国土资源遥感,2012,2(6):28-34
[10] 杜振鹏,李德华．基于KD-Tree搜索和SURF特征的图像匹配算法研究[J]．计算机与数字工程,2012,0(2):96-98
[11] Cornelis N,Gool V L．Fast Scale Invarint Feature Detection and Matching on Programmable Graphics Hardware[C]∥Computer Vision and Pattern Recognition Workshops 2008．AK,2008:1-8
[12] Randima F.GPU Gems:programming techniques,tips and tricks for real-time graphics[M]．Addison-Wesley Professional,2006
[13] GPGPU．General-purpose computation on GPUs．http://www.gpgpu.org,2010-07-04
[14] 钱悦．图形处理器CUDA编程模型的应用研究[J]．计算机与数字工程,2008,36(12):177-180

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed