Detection Of Pupil Fluctuations In A Dark Environment

Dark adaptation is the basic function of the human eye and also very important in visual inspection.Dark adaptation of human relations the ability to work in a dark environment.For pilots, excellent dark adaptation related to the safety of aircraft and passengers.Therefore, the detection of dark adaptation is particularly important.Commonly used detection method is divided into subjective and objective categories.Subjective detection have some uncontrollable factors,such as deliberately false positives. Objective detection using the human physiological response to analyzing the results,such as nystagmus,pupil fluctuation. In low-light conditions,pupil fluctuation is very weak.For the detection of the pupil edges,Fitting algorithm can not reach good results.Many factors influences the pupil fluctuation,such as human emotions.his article converts pupil image into polar coordinate, uses local threshold accurately detects the pupil edge, introduces periodic light stimulus, converts pupil area into frequency domain from time domain, effectively detect fluctuations. Experiment results affirm that the method introduced is suitable for detecting pupil fluctuation in dark environment. The main work of this article is as follows:1.This paper first introduces the traditional edge detection method.Analyzed several classical operators,Such as Sobel, Prewitt, Robert, LOG, Canny.2.Analyzed several points of attention in the image acquisition in the dark environment. For example,the choice of light and the lens.After that according to the specific features and the different application requirements,this paper introduced several methods for detecting the edge of pupil.3.Analyzed the hardware structure and the flow of detection.By combining the polar coordinate transformation and multi-threshold,this paper introduced a new method for detecting the pupil edge.Pupil fluctuation is influenced by the motion of human and other factors,this article using frequency domain analysis to avoid the interference of these factors.In the process of experiment, the edge pixels detected by this method are used as input,the effectiveness of the proposed method is verified in this application environment.