Design And Implementation Of Intelligent Robot Identity Recognition System Based On Image Features

With the rise of artificial intelligence and deep learning,biometric identification technology has undoubtedly become the most effective method to solve the problem of computer authentication and identification.At present,most advanced identification systems use single biological feature,but the use of single biological feature will have problems such as information limitation,expression limitation,and immutability limitation,which makes it difficult to obtain better results for the accuracy of the identity recognition system.Although the multi-biological feature recognition method can improve the accuracy of identity recognition to a certain extent,each feature needs to be collected enough for recognition.Therefore,when using biometric identification technology,select appropriate biometrics,and complete the information collection task through a data collection method.In addition,designing an appropriate fusion method can improve the efficiency and accuracy of the identity recognition system in the actual environment.In response to the above-mentioned problems,the thesis designs the use of robots to collect image data,and completes the task of identity recognition by extracting multiple features from static images for multibiological feature recognition.The main research contents of the thesis are as follows:1.Intelligent robot behavior.Taking the robot as the main body,the data collection tasks of identity recognition are collected in the form of robots.At the same time,it can complete the tasks of automatic inspection,pedestrian perception,active photographing and voice prompts based on the results of identity recognition.2.Use images taken by the robot for identification.Since the robot’s camera has a moderate shooting distance,and the pictures taken are of higher resolution,and clear facial and body shape features can be obtained from the images,so face recognition and overall body shape recognition are selected as the method of identification.3.Research on multi-biological feature recognition model.Traditional biometric identification methods based on static images are easily affected by the actual shooting environment.In order to receive more information and improve the accuracy of identification,the thesis chooses person re-identification(Re ID)as the overall pedestrian body shape recognition method.Face recognition and pedestrian overall body shape recognition are used as research methods for multi-biological feature recognition,so that the robot can accurately identify pedestrians under the complex and diverse conditions of taking photos.The thesis uses robots to actively take pictures of people in the real environment,takes the static image data taken by the robot as the research object,so that the model can extract the required biological features from the image,and uses multiple biological feature recognition technology and feature fusion methods to improve the accuracy of identity recognition.Experiments show that the robot can actively complete the task of identity recognition in a complex real environment.