• DocumentCode
    1725570
  • Title

    A Modified Registration Algorithm Using Affined Transform for W-FOV IR System

  • Author

    Yabing, Chen ; Yongzhong, Wang ; Lihui, Sun

  • Author_Institution
    Mech. Eng. Coll., Shijiazhuang
  • fYear
    2007
  • Abstract
    In this paper, an modified registration algorithm is introduced. Image registration is a key technology in small target detection. Considering the special need of W-FOV IR system, who can work all field and all-time, we establish a 2-D affine transform model, the maximum correlation between image sequences as feature. To balance the effectiveness and running time of algorithm, simplified LSM criterion is used to converse six parameters of affined transform to four parameters. Registration window selected can also shield clutter´s influence. Experimental results using real IR sensor image sequences are presented. Simulation and test shows that this model registration precision can reach single pixel level and shield background clutter influence effectively.
  • Keywords
    affine transforms; image registration; image sequences; infrared imaging; least mean squares methods; object detection; 2-D affine transform model; IR sensor image sequences; LSM criterion; W-FOV IR system; image registration; image sequences; modified registration algorithm; registration window; shield background clutter; small target detection; wide field of view IR detection; Image registration; Image sensors; Image sequences; Infrared detectors; Infrared image sensors; Instruments; Least squares methods; Mechanical variables measurement; Sun; Testing; Affined Transform; IR System; Registration; Wide Field of View (W-FOV); least square method (LSM);
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Electronic Measurement and Instruments, 2007. ICEMI '07. 8th International Conference on
  • Conference_Location
    Xi´an
  • Print_ISBN
    978-1-4244-1136-8
  • Electronic_ISBN
    978-1-4244-1136-8
  • Type

    conf

  • DOI
    10.1109/ICEMI.2007.4350754
  • Filename
    4350754