基于梯度方向约束的瞳孔定位算法研究  

Research on pupil localization algorithm based on gradient direction constraint

作　　者：宁小鸽 牟莉[1] Ning Xiaoge;Mu Li(School of Computer Science,Xi’an Polytechnic University,Xi′an 710600,Chana)

机构地区：[1]西安工程大学计算机科学学院,西安710600

出　　处：《国外电子测量技术》2021年第7期115-121,共7页Foreign Electronic Measurement Technology

摘　　要：针对传统的随机圆检测(RCD)瞳孔定位算法存在无效采样的次数过多和边缘点无效计算,导致瞳孔定位精度低、计算时间长等问题,提出了一种基于梯度方向约束的随机圆检测(GRCD)瞳孔定位算法,该算法首先利用自适应boost算法(Adaboost)人脸检测算法和三庭五眼法对人眼进行粗定位,然后利用Canny算法提取人眼区域轮廓。接着,对人眼区域轮廓采用GRCD算法获得人的眼球中心坐标,即瞳孔位置。最后,在同等条件下,通过MATLAB仿真实验对GRCD和RCD算法分别进行检测,实验结果表明,GRCD算法在准确率、检验速率、鲁棒性上都优于RCD算法,因此该算法是精确度高,耗费时间短,鲁棒性强的瞳孔定位算法。For traditional randomized circle detection(RCD) pupil localization algorithm is the number of invalid sampling too much calculation and edge point, causes the pupil orientation problem such as low accuracy, long computing time, this paper proposes a random circle detection based on gradient direction constraint(GRCD) pupil localization algorithm, this algorithm first USES the Adaboost(adaptive boost algorithm) face detection algorithm and three court five coarse location of human eyes, eye location method and then use Canny algorithm to extract the eye contour area. Then, GRCD algorithm is used to obtain the coordinates of the center of the human eye, namely the position of the pupil. Finally, under the same conditions, GRCD and RCD algorithms were tested respectively through MATLAB simulation experiments. The experimental results show that GRCD algorithm is better than RCD algorithm in running time and accuracy, and it is a pupil positioning algorithm with high accuracy, short time consumption and strong robustness.

关 键 词：人眼粗定位 CANNY算法 梯度方向 GRCD算法 瞳孔定位 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]