Research And Implementation Of Virtual And Reality Interaction Technology In Augmented Reality

Augmented reality is a technology that integrates a computer-generated virtual environment with a physical world environment.It allows users to see that virtual objects appear to become part of the real world.The augmented reality interaction technology is based on the augmented reality system.The user can interact with the virtual information registered in the system,and control the virtual objects.Therefore,the interactive technology can improve user experience.With the improvement of hardware performance and development of augmented reality technology,since controlling virtual objects during the process of interaction between real and virtual objects is not satisfied by using traditional interaction methods and the system lacks realism,the AR applications are several limitations.Hence,most researches focus on AR interactive technology.By analyzing the current researches status and discussing a variety of interactive technologies,this thesis proposes augmented reality interaction methods of instructional interaction and multi-channel interaction based on a Kinect sensor.The methods include three interaction modes: interaction based on depth consistency,static gesture instruction and dynamic gesture instruction.These interactive methods have efficient control over virtual objects and improve the stability,real-time and authenticity of augmented reality interactive systems.The main work and contributions of the thesis are as follows:(1)Aiming at the problem of data noise in the depth image acquired by Kinect sensor,an efficient two-time filters algorithm is proposed,which effectively removes the noise such as pixel jitter and pixel holes in the depth image and ensures the real-time performance of the algorithm.The algorithm also provides high quality image data for the recognition of gesture instruction.(2)For the problem of the low recognition rate and the poor real-time performance of static gesture instruction issued by users,this thesis puts forward the method of pre-processing hand image training data,which ensures the integrity of data features and reduces the learning difficulty of neural network model.On this basis,a convolution neural network model is designed,which can identify static gesture instructions efficiently and accurately.(3)To improve the realism and immersion of the augmented reality interaction system,the 3D human body model is reconstructed by using the depth image data and RGB image data.The interaction way with depth consistency is implemented by performing the functions such as virtual reality occlusion and collision.(4)Investigating the advantages and disadvantages of static gesture instruction and dynamic gesture instruction,a collaborative interactive instruction combining two instructive modes is designed to improve the interaction efficiency and the ability to control virtual objects,and to make the interaction ways rich and diverse.