Research On Face Detection And Recognition Based On Low Resolution Video Surveillance

With the development of science and technology,China has entered a period of rapid development,meanwhile,security problems have also come.To maintain social stability and protect people's lives and property safety,the state and some large enterprises have invested huge amounts of funds in various cities nationwide,and established a relatively perfect monitoring system.But due to the limited capacity of storage equipment makes the effective resolution of video equipment monitoring system is not high,or the target person's distance,makes the face part of the pixel,less blurred images,it is difficult to identify.To recognize the human face correctly and find the target face in this case,the research of face detection and recognition based on super-resolution is carried out.The main research contents are as follows.Introduces several common methods of face detection at present,combined with the actual situation of low resolution on the comparison and selection of the face detection algorithm based on the theory of AdaBoost,the realization of face detection in low resolution conditions.The low resolution of image reconstruction algorithm(i.e.super-resolution algorithm)is introduced from two major domains,frequency domain and spatial domain.The airspace method is analyzed in detail,and the typical 4 algorithms are analyzed in principle.By comparing the advantages and disadvantages of several algorithms,which will be highly complementary MAP and POCS algorithms are mixed,and improve to make it better to avoid using hybrid MAP/POCS algorithm,some deficiencies of mutual advantage.Using simulated Lena video frames after degradation,the algorithm is simulated,and the subjective and objective evaluation methods are used to evaluate the reconstruction effect of several algorithms objectively.This paper introduces the source of face recognition,and analyzes the principle and calculation steps of PCA algorithm in detail.At the same time,the inherent defects are analyzed and the 2DPCA algorithm is improved.The principle of 2DPCA face recognition algorithm and implementation steps are described in detail.The ORL and Yale face databases are used to compare the PCA algorithm with the 2DPCA algorithm.Finally,gives the main framework of the low resolution of video monitoring face detection and recognition based on the combination of installed in the laboratory surveillance camera video information acquisition and pre standard face samples and joint experiments on AdaBoost face detection algorithm,the improved hybrid MAP/POCS super resolution algorithm and 2DPCA face recognition algorithm based on Matlab.The detection and recognition of target face in low resolution surveillance video are realized.The experimental results show that the proposed algorithm can detect the human face in a low resolution monitoring environment,and improved the resolution of the human face,and completed the recognition of the target face.