Smart Home Controller Based On Eye-tracking Technique

Eye-tracking is a new style of human-computer interaction.After detecting the state and characteristics of eyes,it can achieve eye-tracking-based control by locating sight line.In recent years,it has become a research hotspot to integrate the eye-tracking technique into the smart home field.The interactive mode based on eye-tracking has the advantages of non-contact and low interference,and it will have a wide application prospect in smart home control field.According to the cornea-pupil reflection principle,a smart home controller based on eye-tracking technique is built in this thesis.The controller is designed combining SOPC and image processing technologies,and FPGA is selected as the core processor.It can control household appliances by movements of users' eyes.In case of algorithm design,the eye recognition algorithm is designed in the first place,which aims at identifying and locating region of eyes in human face infrared image.Secondly,the sight recognition algorithm is conducted.The centers of pupils and infrared spots will be obtained respectively through image segmentation,edge detection and ellipse fitting.After that,the viewpoint mapping algorithm is projected.Position of viewpoint on screen is calculated by establishing the mapping function between the face image coordinate system and the display screen coordinate system.In addition,a blink detection algorithm is also finished in this thesis.It is used to achieve human-computer interaction.The controller is divided into image acquisition module,image display module,communication module and SOPC module are designed in this thesis.The image acquisition module comprises an infrared illumination unit and a camera.The infrared lighting unit provides near-infrared light,and camera provides video stream for the system.The image display module displays viewpoints and interactive interface using a VGA displayer integrating a D/A chip,and SDRAM is chosen as video memory.The communication module connects with FPGA via GPIO,and it sends the control signals using infrared ray.As the core of system,the SOPC module is employed to implement image processing algorithms,which takes the Nios II kernel provided by the Quartus II development environment as the CPU.Test results show that resource utilization rate of FPGA used in the smart home controller based on eye tracking is down to 11% and the accuracy is over 90%.Not only the controller can realize eye tracking and controlling,but also has enough resource for subsequent upgrade and extension.It meets designed requirements of high precision and scalability.