基于OpenCV的鼻尖特征点鼠标实时控制系统实现  

IMPLEMENTATION OF NOSE TIP FEATURE MOUSE REAL-TIME CONTROL SYSTEM BASED ON OPENCV

作　　者：吴博 王彬[1] 翁政魁 熊新[1] 刘辉[1] 

机构地区：[1]昆明理工大学信息工程与自动化学院,云南昆明650500 [2]上海大学通信与信息工程学院,上海200072

出　　处：《计算机应用与软件》2018年第2期212-217,共6页Computer Applications and Software

基　　金：国家自然科学基金项目(61263017)

摘　　要：针对当前基于视线跟踪或眼部特征的机器视觉鼠标控制系统在实际应用中对环境和条件适应性较差的问题,提出将鼻尖作为人脸局部特征点的实时人脸特征跟踪方法。采用一种带预测机制的改进Viola-Jones框架,在帧间使用Kalman滤波算法对下一帧特征点出现位置进行预测,明显提高了特征点跟踪效率。采用VS2012和OpenCV实现了基于该方法的鼠标指针控制系统。实验结果表明,所实现的鼠标控制系统跟踪处理速度相比传统方法提高了40%左右。同时该鼠标控制系统实时跟踪效果快速、稳定、平滑,具有良好的环境适用性和鲁棒性。Aiming at the problem that the current visual mouse control system based on gaze tracking or eye features had poor adaptability to environment and conditions in practical applications,a real-time facial feature tracking method using nose tip as the local feature point of human face was proposed. This method used a modified Viola-Jones framework with a prediction mechanism to predict the position of the next feature point using Kalman filter between frames,which obviously improved the feature point tracking efficiency. Then the mouse pointer control system was achieved by using the VS2012 and OpenCV based on the method. Experimental results showed that the tracking speed of the mouse control system was increased by 40% compared with the traditional method. At the same time, the real-time tracking performance of the mouse control system was fast,stable and smooth,and had good environmental adaptability and robustness.

关 键 词：鼻尖跟踪 鼠标控制系统 特征点跟踪 OPENCV 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]