头部运动与视线追踪数据融合技术的研究

发布时间：2018-03-11 13:08

本文选题：视线追踪　切入点：头部运动　出处：《西安工业大学》2017年硕士论文　论文类型：学位论文

【摘要】：视线追踪技术在人机界面交互、助残、疲劳驾驶检验、航空医学研究等领域有广泛的应用前景。头部运动是人视线注视过程中的重要的伴随行为,对于人类视觉注意有着重要的意义。通过数据融合技术将头部运动和视线跟踪放在一起研究,开拓了视线追踪系统高效性、可靠性、可使用性和功能性提高的新思路。本文针对头部运动与视线追踪数据融合技术的研究,首先改进视线追踪方法和头部运动跟踪法:对于视线跟踪,本文采用非穿戴式宽视野红外相机获取人面部图像,通过Adaboost算法实现面部识别,然后采用ASM算法对获取的面部图像进行实时特征跟踪,并截取眼球区域图像,然后通过霍夫变换算法和图像阈值处理获取并跟踪眼球区域图像中普尔钦斑和瞳孔中心,最后采用改进的四光源透视法,建立四光源普尔钦斑与瞳孔关系模型,通过多项式估计获得视线注视点。对于头部运动跟踪,本文同时采用基于惯性传感器和基于POSIT摄像机观测法的头部姿态空间运动的跟踪。首先本文通过四阶龙格库塔法、毕卡逼近和加速度插值积分,实现惯性传感器的姿态运动跟踪。然后通过POSIT迭代算法对图像进行解算获得头部运动信息。最后结合两种方法的特点,通过数据融合,取长补短,有效提高头部运动跟踪测量效率。本文对数据融合技术进行理论研究,提出了数据融合机的概念,以及结合神经网络算法等智能方法实现数据融合系统的构建。基于理论研究,将获取的眼动数据和头部运动数据进行融合处理,从而获得视线跟踪系统的头部补偿与视线空间方向数据,为头眼协调等应用提供测量工具。最后,本文结合软件体系结构,实现了非穿戴式的头部运动与视线追踪数据融合测量系统。系统由多个软件构件构成,输入数据通过连接件传入构件,输出所需要的数据,构成一套具备完整功能的软件。
[Abstract]:Eye tracking technology has a wide application prospect in the fields of man-machine interface interaction, disability, fatigue driving test, aeronautical medical research, etc. Head movement is an important accompanying behavior in the process of human visual gaze. It is of great significance to human visual attention. Through data fusion technology, the head motion and line of sight tracking are studied together to develop the high efficiency and reliability of the line of sight tracking system. This paper aims at the research of head motion and line of sight tracking data fusion technology, first improve the line of sight tracking method and head motion tracking method: for line of sight tracking, In this paper, a non-wearable wide-field infrared camera is used to obtain human facial image, and Adaboost algorithm is used to realize facial recognition. Then, ASM algorithm is used to track the real-time features of the acquired facial image and to capture the eyeball region image. Then, by means of Hough transform algorithm and image threshold processing, Pulchin spot and pupil center in the eye region image are obtained and tracked. Finally, an improved four-light source perspective method is used to establish the model of the relationship between the four light source Purchin spot and the pupil. For head motion tracking, both inertial sensor and POSIT camera observation method are used to track the head attitude spatial motion. Firstly, the fourth order Runge-Kutta method is used in this paper. The attitude tracking of inertial sensor is realized by using Bika approximation and acceleration interpolation integral. Then the head motion information is obtained by POSIT iterative algorithm. Finally, combining the characteristics of the two methods, we learn from each other by data fusion. In this paper, the data fusion technology is studied theoretically, the concept of data fusion machine is put forward, and the intelligent method such as neural network algorithm is used to realize the construction of data fusion system. The obtained eye movement data and head motion data are fused to obtain the head compensation and line of sight spatial direction data of the line of sight tracking system, which provides a measurement tool for head-eye coordination and other applications. Finally, this paper combines the software architecture, A non-wearable head motion and line of sight tracking data fusion measurement system is implemented. The system is composed of several software components. The input data is transmitted to the components through the connectors and the required data is outputted to form a set of software with complete functions.
【学位授予单位】：西安工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41

【相似文献】