基于深度图像信息的指尖跟踪与轨迹识别  

FINGERTIP TRACKING AND TRAJECTORY RECOGNITION BASED ON DEPTH IMAGE INFORMATION

作　　者：李哲[1] 彭四伟[1] 

机构地区：[1]北京化工大学信息科学与技术学院,北京100029

出　　处：《计算机应用与软件》2016年第4期155-159,172,共6页Computer Applications and Software

摘　　要：针对基于传统摄像头获取的二维图像难以准确区分复杂环境下目标和背景的问题,提出一种利用Kinect摄像头对复杂背景下手指指尖的跟踪算法,并对指尖跟踪的轨迹进行识别。首先利用Kinect获取的深度图像信息对手部进行分割和指尖检测;然后利用压缩感知方法对跟踪目标进行特征提取,通过朴素贝叶斯(NB)分类器分类跟踪目标和背景;最后,通过支持向量机(SVM)方法对跟踪的轨迹进行识别。实验结果表明,提出的方法能够成功地跟踪手指指尖的位置,同时能够准确识别出指尖跟踪的轨迹。It is difficult to accurately distinguish the object and the background under the complex environment in two-dimensional image acquired based on traditional camera. To solve the problem,we proposed a tracking method which utilises Kinect camera aiming at the fingertip in complex background,and recognises the trajectory of fingertip tracking. First,it uses the depth image information captured by Kinect to segment the hand and to detect the fingertip. Subsequently,it uses compressive sensing method to extract features from tracking object,and employs the naive Bayesian（ NB） algorithm to classify the tracking object and background. Finally,it uses support vector machine（ SVM） method to recognise the tracking trajectory. Experimental results showed that the proposed method can successfully track the fingertip locations and accurately recognise the fingertip trajectory at the same time.

关 键 词：KINECT 指尖跟踪 深度图像 压缩感知 手势轨迹识别 支持向量机 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]