Real-time measurement method of pupil diameters based on binocular stereo vision

Abstract

A real-time measurement method of pupil diameters based on the binocular stereo vision is provided. The method comprises: using Zhang Zhengyou's dual target method for camera calibration to obtain a rotation matrix, a translation matrix, a parameter matrix in the left industrial camera, and a parameter matrix in the right industrial camera; respectively processing images captured by the left industrial camera and the right industrial camera to obtain true ellipses corresponding to the pupil; drawing a straight line across the center of the true ellipse of the pupil obtained by the left industrial camera, wherein the straight line equation is intersected by the true ellipse of the left industrial camera to obtain a first intersection point and a second intersection point; according to the principle of polar line constraint and order consistency, by using the first intersection point and the second intersection point, respectively solving a first matching point and a second matching point corresponding to the two intersection points on the true ellipse obtained by the right industrial camera; and obtaining preliminary values of a plurality of pupil diameters corresponding to a plurality of straight lines with different slopes, and solving an average value to obtain a final result. The method provided by the present invention has high measurement precision, does not require the worker to fix the working distance, and has great application value.

Description

technical field [0001] The invention relates to a method for measuring pupil diameter in real time. In particular, it relates to a real-time measurement method of pupil diameter based on binocular stereo vision. Background technique [0002] Pupil oscillations contain a lot of important information. In recent years, many research institutions have researched and developed various systems that utilize pupil oscillation information. Pupil oscillations can be used to study human emotions, such as seeing pleasant or unpleasant images, feeling to irritability, stress, and reactions when seeing exciting products; some research institutions study the relationship between pupil oscillations and driver fatigue; in the medical field, pupil oscillations can also be used to support doctors in the detection of early diabetes. [0003] Pupillometry based on computer vision can use visible light or infrared light as a light source; although visible light is a more convenient choice, the p...

AI Technical Summary

Problems solved by technology

[0007] Recently, several pupil center and edge detection algorithms have been proposed, such as ElSe (Conference: Biennial ACM Symposium on Eye Tracking Research & Applications; Authors: W. Fuhl, T.C. Santini and E. Kasneci; Year of publication: 2016; Article title: ElSe: ellipse selection for robust pupil detection in real-world environments; pages: 123-130), ExCuSe (Conference: Computer Analysis of Images and Patterns; Authors: W.Fuhl, T.Kübler, K.Sippel, W.Rosenstieland E.Kasneci; year of publication Month: 2015; Article Title: ExCuSe: robust pupil detection in real-world scenarios; Pages: 39-51) and Swirski (Conference: Proceedings of the Symposium on EyeTracking Research and Applications; Authors: L.Swirski, A.Bulling, and N.Dodgson; publication year: 2012; article title: Robust real-time pupil tracking in highly off-axis images, page number: 173-176), etc., most of these algorithms are developed for head-mounted systems; head-mounted The traditional system can generally obtain high-resolution graphics for the human eye area, but the low-resolution images of the human eye area usually obtained by the remote device are low; in addition to image quality, the remote system also needs to deal with extreme head angles The resulting camera angle of view is insufficient or the eyes are not visible in the image; therefore, in practical situations, these algorithms need to be adjusted for long-distance applications (Conference: Proceedings of the Symposium on Eye Tracking Research and Applications; Author: L. Swirski ,A. Bulling and N.Dodgson; Year of publication: 2012; Article title: Robust real-time pupil tracking in highly off-axis images; Page number: 173-176)

Images