Research On Multi-interaction System Of Virtual Prototype Based On Augmented Reality

In recent years,with the development of virtual reality and augmented reality technology,its application areas have become more and more widespread.At the same time,the real estate industry has been a very popular industry in all industries,and its promotional marketing methods have also created innovative demands along with the development of science and technology.The application of augmented reality technology to the real estate industry can not only open up the application fields of high-end technologies,carry out new application development,but also apply science and technology to real estate,and is also a breakthrough innovation development in terms of sales methods for real estate.Aiming at the innovative demand of the real estate industry's sales means,a 3D model of the template is created based on 3DMAX.Based on the Unreal Engine4 virtual reality development engine,a virtual prototype multi interaction system based on augmented reality is developed,which realizes the roaming,video embedding,automatic induction equipment simulation,replacement model,material,free moving model,and the model of the multi interaction between the three interactive modes of voice control,gesture control and hardware device control.Voice activated lights and their specific functions such as conversion from two-dimensional code to mobile client and other functions have changed the single way of human-machine interaction in the past.The speech recognition uses the first order high pass filter to preprocess the speech,uses the method of Mel frequency cepstrum coefficient(MFCC)to extract the feature,uses the mixed Gauss model and the hidden Markov model as the acoustic model,uses the N-Gram model to train the language model,and finally uses the Viterbi algorithm to realize the speech.Decoding and searching,through the recognition of the key words and functional functions,and other steps to achieve the "open light","close" and other keyword speech recognition and control virtual space model operation.The gesture control is mainly based on the HSV skin model and morphological operation,by intercepting the frame image of the visual frequency flow as input,converted to HSV space,morphological processing is used to extract the outline of hand,and the gesture recognition and control is realized through the change of the position and shape of the hand contour.The whole process mainly includes motion detection,image preprocessing,image segmentation,morphological processing,contour acquisition,information comparison and vector calibration,so as to achieve dynamic gestures.Single identification,based on the above algorithm,realizes the control of multiple interaction means.