| With the rapid development of artificial intelligence technology,and the automobile industry has a new development direction.At present,applying this technology to the traditional automotive industry has become the research focus of universities and institutions.The application of face recognition technology in the intelligent vehicle platform can not only enable the face to start the vehicle and effectively reduce the possibility of illegal driving of the vehicle,but also open various personalized services customized by users according to the identification results to provide users with intelligent driving experience.In this thesis,face recognition and OCR technology are studied deeply.In order to adapt to the application environment,the traditional face recognition algorithm is improved according to the algorithm with advantages in OCR.(1)The combination of traditional face detection algorithm and character detection algorithm based on connected domain analysis.The binary image is obtained by clustering the skin color,then using connected region analysis algorithm and Ada Boost algorithm to mark,screen and verify the candidate region,and get the face region.(2)Research and analyze a variety of preprocessing algorithms,and select the appropriate algorithm.In the coarse positioning area,the histogram projection method is used to quickly and accurately position the eyes,and the angle and scale parameters used for rotation correction are obtained.(3)References are made to the advantageous algorithms and ideas in OCR technology.The LTP features are extracted on the basis of key points,and the division method uses a non-uniform grid.The whole similarity is calculated by weighted method to realize matching classification.The simulation results show that the optimized face recognition algorithm has better recognition rate and speed,and can better meet the requirements of the intelligent vehicle Platform.It has certain practical significance and practical value. |
PDF Full Text Request