This paper presents a simple and effective pupil localization method based on binary threshold coarse location and region growing.The morphology is used as a pretreatment method to remove the influence of the eyelash and the light spot,which could keep the original pupil size and generate thumbnails at the same rime.To find the extreme point of the binary image,which can be used as the seed point,the horizontal and vertical projection is conducted.Region growing algorithm is used to obtain the accurate pupil region.The results demonstrate that the pupil parameters could be detected in real-time by using the acquisition system which is developed using OpenCV technology,the processing rate of the system can be at 29.75 fps.%提出一种基于二值化粗定位和区域生长方法相结合的方法,能有效检测因病变而呈不规则圆形的瞳孔,简单而且快速.首先使用形态学开运算进行预处理,去除睫毛和亮斑的影响且保持原有瞳孔大小,同时生成缩略图;然后生成二值阈值化图像,进行水平和垂直投影,找到极值点,将其作为种子点;最后用区域生长算法得到精确的瞳孔区域.试验结果显示:用OpenCV构建的采集系统能够实时地对瞳孔参数进行检测,达到29.75帧/秒.
展开▼
