The SURF algorithm on a FPGA-based platform

The SURF algorithm is suitable for being implemented on small computation platforms for its advantages of scale-invariant, rotation-invariant, and low computation burden. This paper focuses on the realization of the SURF algorithm upon a small FPGA-based platform for objects recognition in real-time. By following the hardware and software co-design concept the entire system is implemented and embedded in the FPGA of a SoPC. The hardware part of the system includes image processing circuits and peripheral circuits. The image processing circuits include RGB to gray transformation circuit and image integral circuit. Peripheral circuits include Laplacian of Gaussian filters circuit and NiosII process circuit. In software part, it includes SURF related operations such as: interest point descriptor, interest point matching, and the software for the system control. It has experimentally been proven that the resulting system satisfies the design goals of objects recognition, rotation-invariant, and scale-invariant. But, in the resulting system, there are only several modules realized in hardware because of the hardware resource limitation of the FPGA chip, causing the recognition speed cannot have very much improvement.

[1]  Hossein Nezamabadi-pour,et al.  Concept learning by fuzzy k-NN classification and relevance feedback for efficient image retrieval , 2009, Expert Syst. Appl..

[2]  Cordelia Schmid,et al.  Scale & Affine Invariant Interest Point Detectors , 2004, International Journal of Computer Vision.

[3]  Jae Wook Jeon,et al.  FPGA-based real-time visual tracking system using adaptive color histograms , 2007, 2007 IEEE International Conference on Robotics and Biomimetics (ROBIO).

[4]  Xiangyang Xue,et al.  Content and context-based multi-label image annotation , 2009, 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops.

[5]  Berthold K. P. Horn,et al.  "Determining optical flow": A Retrospective , 1993, Artif. Intell..

[6]  Cordelia Schmid,et al.  An Affine Invariant Interest Point Detector , 2002, ECCV.

[7]  Hans P. Moravec.  Obstacle avoidance and navigation in the real world by a seeing robot rover , 1980 .

[8]  T. Lindeberg,et al.  Scale-Space Theory : A Basic Tool for Analysing Structures at Different Scales , 1994 .

[9]  Christopher G. Harris,et al.  A Combined Corner and Edge Detector , 1988, Alvey Vision Conference.

[10]  Matthijs C. Dorst.  Distinctive Image Features from Scale-Invariant Keypoints , 2011 .

[11]  Berthold K. P. Horn,et al.  Determining Optical Flow , 1981, Other Conferences.

[12]  Matthew A. Brown,et al.  Invariant Features from Interest Point Groups , 2002, BMVC.

[13]  Avinash C. Kak,et al.  Vision for Mobile Robot Navigation: A Survey , 2002, IEEE Trans. Pattern Anal. Mach. Intell..

[14]  Luc Van Gool,et al.  SURF: Speeded Up Robust Features , 2006, ECCV.

[15]  Michel Dhome,et al.  Real Time Localization and 3D Reconstruction , 2006, 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06).