Research On The Design Of Portrait Acquisition And Intelligent Identification System

With the advent of the post-PC era,intelligent terminal devices,which combine modern computer technology represented by embedded technology with consumer products,are constantly appearing in family life,and can efficiently accomplish specific tasks.For the application requirements of home security systems,this paper researches and designs a set of face image acquisition and intelligent identification system based on ARM platform.Hardware modules,software systems,and function implementation modules are consisted in the system.The software system consists of three parts: startup program,kernel,and file system.Among them,Vboot is used as the startup program,which supports NandFlash boot mode,and completes the system initialization work;Tailored kernel configured system hardware drivers and root file types,effectively manages system resources;By designing its file content,the root file system matched the kernel and startup programs ensures the orderly operation of the system.System function is achieved by four modules,which are image acquisition,preprocessing,face detection and face recognition.The frame images that the camera mapped to the memory are extracted by the image acquisition module and converted channel to solve the problem of mismatch between the window display and the camera.The histogram equalization and filter are used to improve the picture's quality in the preprocessing module and which guarantees the effectiveness of feature extraction.The face detection module is designed to determine whether the face part is included in the picture through the Adaboost cascade classifier whose inputs are Haar features.In this process,the detection accuracy is improved by adding human eye positioning detection;LBP algorithm is used in the Face recognition module to extract face features,and the space conversion method is used to reduce the data dimensionality,which effectively reducing the amount of calculation,and then the face recognition is completed by the classifier through the data comparison.Finally according to the needs of the application,the system interaction interface is designed and the images can be displayed on the window.Based on the above studies,the hardware system was built,the software system and the function module's program are transplanted to the ARM,and the system was experimentally verified.The verification results show that the system has the function of detecting and recognizing the faces in the images captured by the camera.In the case of ensuring accurate detection of human face and human eyes,the detection time is about 100 ms,the recognition time is about 0.5s,and the recognition rate is over 90% under good light conditions,The realization of the system in this paper proves the feasibility of the embedded intelligent monitoring equipment and has great reference value for the perfection and promotion of the intelligent monitoring system.