A Super-resolution Algorithm Based on SURF and POCS for 3D Bionics PTZ

Sensors & Transducers

Jiqing CHEN¹  Shaorong XIE¹  Jun LUO¹  Hengyu LI¹ 

[1] School of Mechatronics Engineering and Automation, Shanghai University, 149 Yanchang Road, Zhabei District, Shanghai, 200072, China;

关键词: Super-resolution Algorithm;    SURF;    POCS;    Bionics PTZ;    RANdom SAmple Consensus.;   

DOI  :  

来源: DOAJ

【 摘 要 】

Image super-resolution algorithm improves image resolution in software. Before image tracking, it needs to enhance image resolution to improve the image tracking accuracy of the bionic eye. The traditional super-resolution reconstruction method can’t satisfy the accuracy and real-time system, so this paper proposes super-resolution image reconstruction algorithm based on SURF (Speeded up Robust Features) and POCS (Projections Onto Convex Sets). The algorithm applies SURF algorithm on image registration and uses RANSAC (RANdom SAmple Consensus) algorithm to kick out fault feature to improve the accuracy of image registration. After motion estimation, this paper applies POCS algorithm to reconstruct a super-resolution image. Finally, some experiments are performed on the platform of a bionic eye, which show that the algorithm can better improve the image resolution and reach real-time requirements to some extent, and providing basis for the subsequent image tracking.

【 授权许可】

Unknown