Title :
A New Horizon Electronic Image Stabilization Algorithm Based on SVM
Author :
Yantai, Xue ; Tao, Wu
Author_Institution :
Coll. of Mechatron. Eng. & Autom., Nat. Univ. of Defence Technol., Changsha, China
Abstract :
In this paper, SVM is used in an electronic image stablizing algorithm designed for the sea environment with the horizon feature. First, samples updated mechanism proposed makes SVM an online classifier to distinguish two regions. On this basis, through the binary image processing and the Hough transform, the horizon is detected and the motion compensation values are acquired. Finally, the Affine transform and Bilinear interpolation is adopted to rectify the image sequence. These strategies make the algorithm have advantages in both high accuracy and low computational complexity.
Keywords :
Hough transforms; affine transforms; computational complexity; image processing; interpolation; pattern classification; support vector machines; Affine transform; Hough transform; SVM; bilinear interpolation; binary image processing; computational complexity; horizon electronic image stabilization algorithm; online classifier; sea environment; electronic image stabilization; image processing; samples updated mechanism; support vector machine;
Conference_Titel :
Optoelectronics and Image Processing (ICOIP), 2010 International Conference on
Conference_Location :
Haiko
Print_ISBN :
978-1-4244-8683-0
DOI :
10.1109/ICOIP.2010.343