Research On Interactive Shadow Puppet Robot Based On Machine Vision

With the development of the social era,people's lifestyles have undergone earth-shaking changes,and entertainment products based on modern technologies have fully entered all walks of life.However,many folk traditional culture represented by shadow play have gradually withdrawn from the historical stage.From the perspective of inheriting and developing shadow play and carrying forward the excellent traditional Chinese culture,in order to allow the public to better accept and enjoy shadow play culture,machine vision technology is applied to shadow play performance.A parallel interactive shadow puppet robot with face detection,hand recognition and face tracking is designed.In the scheme,face detection and face tracking algorithm are improved and optimized.This paper determines the visual interaction function and proposes the system design scheme.The machine vision function is realized by the computer system.The shadow robot adopts the parallel robot structure,and uses the computer system and the STM32 single-chip joint design program to realize the interactive motion control.Firstly,the acquired image is denoised by gray level in the face detection stage,and the image is preprocessed by Gamma correction and wavelet transform.Combined with the traditional MB-LBP,the PMB-LBP is proposed.By combining circular LBP operator and PMB-LBP operator,multi-feature fusion texture features of wavelet image are obtained,and then the SVM is used to train the image features to achieve face detection with high accuracy.Secondly,in the phase of face tracking,an environment adaptive improved tracking algorithm is proposed.The Camshift algorithm is used to track the detected face,and real-time discrimination of the tracking environment is carried out.Linear prediction mechanism is adopted for face blocking,background interference and other situations.In the case of good tracking,Kalman prediction mechanism is adopted to achieve face tracking with high accuracy,which provides more accurate data for the movement of the shadow robot.Finally,the accuracy of face detection,human recognition and face tracking is verified by building a computer system vision and interactive execution platform.Experimental results show that the visual interactive shadow puppet robot can achieve very good effect on human-computer interaction;the scheme of theory analysis is consistent with the experiment results,which proves the correctness of the design analysis.As a visual interactive project,this system has a good cultural heritage,and it has certain reference value for the application of visual interactive robot in promoting traditional culture.