| Eye tracker is professional equipment designed to test and record the changes of eyes, and thus follow the line of sight. It is the product of eye tracking techniques. Eye tracker is of great significance in human cognition, psychology, industrial engineering, computer science, and many other areas.Eye tracking technology can be divided into two aspects:theoretical research and system implementation. This thesis focuses on head-mounted eye tracker's calibration and implementation.This thesis discusses the implementation framework of head-mounted eye tracker based on the general overview of eye tracking technology, and gives functional description on eye image processing module, module calibration, and optical subsystem.Eye tracker optical subsystem which is designed by corneal-pupil reflection method chooses infrared light of 860nm wavelength based on physiological characteristics of eyes. This thesis detailed descripes the geometric position relations of eye camera and scene camera in the design of helmet according to relationship between reflection imaging spot, principle of coordinate transformation and principle of camera calibration.When designing the calibration system, we design two calibration programs according to physical model of eyeballs. The first program to simulate the relationship between eye observation position and pupil and corneal glint based on physiological characteristics and optical principle of eyes after having a clear understanding of light of sight and optical axis. The second program is to obtain calibration relationship according to the fitting of simulation curve of formulas, the relationship between eye position and fixation derived from geometric relations and the simplified model of eyes. Finally, using MFC-based software systems, we verify calibration algorithm through static calibration which uses computer screen as observation field and dynamic calibration which uses realistic scenario as observation field repectively. |
PDF Full Text Request