基于瞳孔定位的眼底相机照明控制技术  

Fundus Camera Illumination Control Technology Based on Pupillocation

作　　者：赵宝彬 刘雨晴 郝如龙 葛爱明[1] ZHAO Baobin;LIU Yuqing;HAO Rulong;GE Aiming(Department of Light Sources and Illuminating Engineering,Fudan University,Shanghai 200433,China)

机构地区：[1]复旦大学光源与照明工程系,上海200433

出　　处：《照明工程学报》2020年第6期158-163,174,共7页China Illuminating Engineering Journal

摘　　要：为辅助眼底相机对患者自动进行眼底成像,需要眼底相机实时检测患者瞳孔中心,并及时触发眼底相机的照明控制系统,使得照明光束及时准确射入眼底。首先,使用本文提出的基于内部集成ARM+FPGA结构的ZYNQ处理器的瞳孔中心点快速定位技术完成患者瞳孔中心点坐标计算;然后,计算患者瞳孔中心点坐标与眼底相机成像视场中心点坐标的距离;最后,当检测到该距离小于触发阈值时,即患者瞳孔已经处成像视场中心,此时立刻触发照明系统完成照明动作,使得照明光束能够精准通过瞳孔照亮眼底。实验结果表明:该系统可在瞳孔部分被遮挡(眼睑遮挡或眼球转动到眼角处)、眼动和眼睫毛背景干扰以及离焦等多种情况下,快速准确计算出瞳孔中心点坐标,并且正确判断是否满足照明条件,瞳孔定位的准确率为96%,检测时间每帧69 ms,基本满足眼底相机对基于瞳孔定位的照明控制技术的鲁棒性、精确性和实时性的要求,有利于眼底相机的智能化改进。In order to assist the fundus camera to automatically collect fundus image of the patient,the fundus camera is required to complete the pupil center location and trigger the illumination system of the fundus camera in time,so that the illumination beam can be timely and accurately injected into the fundus.First,the pupil center location based on the ZYNQ processor integrated with ARM and FPGA proposed in this paper is used to calculate the coordinate of the patient’s pupil center;Next,the distance between the coordinates of the patient’s pupil center and the imaging field of view of the fundus camera is calculated;Finally,when it is detected that the distance is less than the trigger threshold,that is,the patient’s pupil is already at the center of the imaging field of view,the light source switch is immediately triggered to perform the illumination action,so that the illumination beam can accurately illuminates the fundus through the pupil.Experimental results indicate that the system can quickly and accurately calculate the coordinates of the pupil center point and correctly determine whether the lighting conditions are met in a variety of situations,such as obstruction of the pupil(eyelid occlusion or eyeball rotation to the corner of the eye),eye movement,eyelash background interference,and defocus.The accuracy rate is 96%,and the detection time is 69 ms per frame.The proposed method meet the requirements of robustness,accuracy and real-time for lighting control technology based on pupil center location,which is conducive to the development and promotion of new smart fundus camera.

关 键 词：眼底相机 瞳孔识别 ZYNQ 照明 

分 类 号：TP394.1[自动化与计算机技术—计算机应用技术]