Research And Implementation Of Service Robot Based On Speech Interaction And Face Recognition

With the trend of technological change,people's demand for service robots has become more and more intense,and their industrial development is in the ascendant.Voice interaction and face recognition are the basic functions of service robots.In the working scene of the service robot,the technical bottleneck of the robot and the uncertain factors in the environment will affect the user's interactive experience.In the speech interaction,some service robots have large errors in the recognition results due to technical defects.Even if the recognition is correct,the answers are wrong owing to the lack of matching strategies.In face recognition,some service robots are very sensitive to illumination variation and random noise,and the recognition result is highly affected.At the same time,some face recognition algorithms have defects,and their accuracy is low in the case of multi-face existence and face pose transformation.Aiming at the problems of speech interaction and face recognition in service robots,this paper combines the latest research results such as current pattern recognition.And the paper realizes the text recognition of user's questioning voice by means of speech recognition technology.According to the TF-IDF algorithm,the core keywords in the question can be extracted and the targeted answers can be given by keywords.By using HOG features and feature descriptors,users can be identified in complex work scenarios.The anti-interference is good and accuracy of face recognition is high.The application scenario takes the library service robot as an example,and a service robot software system is designed.The system provides users with intelligent services for library applications.The main work and results of this paper are as follows:Firstly,the voice interaction function of the service robot is realized.First of all,the system performs noise reduction processing on the user's question,and calls the voice recognition SDK to identify it.Then the TF-IDF algorithm is selected to extract the keywords in the question.Finally,the similarity comparison is made with the problems in the corpus,and the answer with the highest matching degree is obtained.Secondly,the face recognition function of the service robot is realized.In the first place,the system performs illumination compensation and image denoising on the video frame,and it selects the HOG feature to detect the face.Afterwards,the feature descriptor is extracted by Res Net,and the target face is locked by the face tracking algorithm combined with the feature descriptor.Finally,according to the feature description of the target face,the result is compared with the face database to discern the identity.Thirdly,by taking a library service robot as an example,a service robot software system is realized.The system can realize the voice question and answer function between the service robot and the user.It can accurately recognize the face in the real scene and record the identification information,thereby reducing the workload of the manager.At the same time,managers can freely operate the corpus and face database in the background management,so that the contents of the library can be updated and adjusted in time.The implementation of the service robot software system based on voice interaction and face recognition proposed in this paper meets the actual needs of the current service robot.The software system was put into trial use in a certain university library with good results.