Design And Implementation Of Face Recognition System For The Guest Welcoming Robot

With the development of science and technology,robots are playing an increasingly important role in human life.Service robot can do some repetitive and tedious work for human beings,which not only saves human labor,but also further promotes the development of society.Among them,the guest welcoming robot has a very broad market prospect with good business greeting,human-computer interaction and other functions.As an important part of the vision system of the guest welcoming robot,face recognition is a key technology to realize the human-computer interaction of the guest welcoming robot.As a biometric identification technology based on human facial feature information,face recognition has a very broad application prospect with its advantages of non-intrusive,convenience,friendliness,non-contact and extensibility.This paper proposes the method of using cloud platform as the "brain" of the guest welcoming robot,which solves the problem that most of the guest welcoming robots use the built-in computer for face recognition and thus lack computing power.Considering the complex and changeable light in the working environment of the guest welcoming robot,a face detection method incorporating Ada Boost and Retinex is proposed,which improves the accuracy of face detection under complex light environment.The face recognition method based on deep learning is used to replace the traditional face recognition method,which greatly improves the accuracy of face recognition.Finally,the face recognition system is implemented on the guest welcoming robot platform and the relevant functions are tested.It has good practicability.The main work of this paper is divided into the following three aspects:(1)Aiming at the disadvantage that the face recognition of robots on the market mainly relies on the built-in computer,which leads to insufficient computing power,this paper puts forward that the cloud platform is used as the "brain" of the guest welcoming robot,and the guest welcoming robot is used as the cloud robot to communicate with the cloud platform through Wi Fi.The human face detected by the guest welcoming robot is uploaded to the cloud platform for face recognition calculation,which reduces the computational load of the local robot.At the same time,the cloud platform can also provide cloud services for multiple guest welcoming robots.(2)This paper focuses on the face detection algorithm based on Ada Boost.Considering the advantages of Ada Boost algorithm that is simple and fast,the face detection module based on Ada Boost is designed by using Open CV.Considering the complex and changeable light in the working environment of the guest welcoming robot,a modified multi-scale Retinex algorithm is combined,which further improves the accuracy of face detection.On the cloud server,the Tensorflow deep learning framework is used as the development platform,and the Face Net network model is used to design the face recognition module.Compared with the traditional face recognition method,the accuracy is greatly improved.(3)A face recognition system is designed and implemented on the guest welcoming robot platform,and its functions are tested.At the same time,the performance of the whole system is tested.The system is optimized according to the test data,which improves the performance of the system and makes it practical.