The Research And Design Of Anti Light Interference Face Recognition Algorithm In The Smart Home System

Face recognition system has the characteristics of user-friendliness,automatic recognition,but the recognition rate will descend under flexible illumination.Since complex environment influences the recognition rate greatly,the research on how to increase recognition rate of face recognition systems has important value.In this thesis,we mainly discuss how to improve the recognition rate of face recognition system in complex illumination environment.The traditional method to solve the illumination problem is to solve the influence of illumination through handling the face image based on illumination compensation.However,methods based on illumination compensation weakened the illumination influence to a certain extent.In some special occasions,such as Smart Home scene which is under the large influence of illumination,using these methods to solve the problem becomes difficult,hence,we must study a new algorithm to solve under the specific scene illumination problem.Smart Home demand that the recognition rate of face recognition algorithm should be as high as it can.With a view to characteristics of smart Home Furnishing,this theses proposed the improved HV2 DPCA algorithm and illumination equalization preprocessing transformation.These two methods which exist in face image preprocessing stage and face recognition stage relieve low recognition rate problems caused by illumination in Smart Home scene,and significantly improve the recognition accuracy.In order to illustrate clearly the promotion of the improved HV2 DPCA algorithm and illumination equalization preprocessing transformation,this thesis developed a set of face recognition simulation system based on MATLAB,using this system to get data which is obtained before and after the use of anti light interference method,and then we can draw the curve graph using these data,Finally we can carry out corresponding analysis by the graph.In order to verify the fact that the improved algorithm perform well in the Smart Home scene,we developed a face recognition system and embedded it into the ARM board,then through the simulation of harsh light conditions in the Smart Home scene,we obtained face recognition rate data before and after the improved algorithm,finally we can conclude that the improved algorithm improve the recognition rate.