Date: September 16, 2005

Time: 3:00 p.m.

Location: GCATT Room 325

Speaker(s): Dalong Li

Title: A New Method for Suppressing Optical Turbulence in Video

Abstract:
The presence of optical turbulence can contribute significantly to degradation in video. This problem arises routinely, for example, in astronomy where objects reside beyond the earth's atmosphere. A new technique for suppressing turbulence is described. We consider both the blurring and the geometric distortion when modeling the degradation of atmospheric turbulence. Kurtosis minimization is used to identify the blur. Then Wiener filter performs the deblurring. Geometric distortion is handled through motion compensation. Motion is computed based on adaptive control grid interpolation. The quasi-periodicities of the turbulent motion are exploited, which allow them to be suppressed while true motions of camera (such as panning and zooming) and objects are preserved. The technique is believed to be beneficial in astronomy, atmospheric physics, and surveillance.

Bio:
Dalong Li came to Georgia Tech for his PhD program in electrical engineering in 2001 after he received his master degree in Chinese Academy of Sciences. His research interests are in computer vision, pattern recognition and image processing. Over the years he has interned for HP, Kodak, the Mathworks, and Philips.

Date:	September 16, 2005
Time:	3:00 p.m.
Location:	GCATT Room 325
Speaker(s):	Dalong Li
Title:	A New Method for Suppressing Optical Turbulence in Video
Abstract:	The presence of optical turbulence can contribute significantly to degradation in video. This problem arises routinely, for example, in astronomy where objects reside beyond the earth's atmosphere. A new technique for suppressing turbulence is described. We consider both the blurring and the geometric distortion when modeling the degradation of atmospheric turbulence. Kurtosis minimization is used to identify the blur. Then Wiener filter performs the deblurring. Geometric distortion is handled through motion compensation. Motion is computed based on adaptive control grid interpolation. The quasi-periodicities of the turbulent motion are exploited, which allow them to be suppressed while true motions of camera (such as panning and zooming) and objects are preserved. The technique is believed to be beneficial in astronomy, atmospheric physics, and surveillance.
Bio:	Dalong Li came to Georgia Tech for his PhD program in electrical engineering in 2001 after he received his master degree in Chinese Academy of Sciences. His research interests are in computer vision, pattern recognition and image processing. Over the years he has interned for HP, Kodak, the Mathworks, and Philips.