Implementation of Image Registration Algorithms for Real-time Target Tracking Through Video Sequences

Defence Science Journal

Y. Dilip¹  Jharna Majumdar¹ 

[1]Aeronautical Development Establishment, Bangalore

关键词: Umnanned aerial vehicle;    image registration;    misregishation;    real-time target backing;    target occlusion;    cubic-spline model;    Kalman filter;   

DOI  :  

学科分类：社会科学、人文和艺术（综合）

来源: Defence Scientific Information & Documentation Centre

【 摘 要 】

"Automatic detection and tracking ofinteresting targets from a sequence of images obtained from areconnaissance platform is an interesting area of research fordefence-related applications. Image registration is the basic step usedin target tracking application. The paper briefly reviews some of theimage registration algorithms, analyse their performance using asuitable image processing hardware, and selects the most suitablealgorithm for a real-time target tracking application using cubic-splinemodel and spline model Kalman filter for the prediction of an occludedtarget. The algorithms developed are implemented in a ground-based imageexploitation system (GIES) developed at the Aeronautical DevelopmentEstablishment for unmanned aerial vehicle application, and the resultspresented for the images obtained during actual flight trial.

【 授权许可】

Unknown