The Human Eye To Identify Its DaVinci System Implementation

The eyes recognition is the use of image processing technology to detect region and identify the eye location and status of the human eye in the video stream. Important as the field of computer image processing and pattern recognition research, the theory of the eyes identify algorithm for further studies, fatigue driving, drunk driving detection and authentication of modern transportation and security and other fields increasingly in-depth application practice. The thesis is in-depth study of the eye detection, identification and localization algorithm and its realization Davinci system on the embedded platform, based on the eye recognition.Identification of the human eyes is based on face detection. This paper is based on the methods of color features face detection. Establish skin model and color segment to determine the face region finally. The eyes recognition for the determination of the face region is the use of integral projection method to determine the eye area. Finally the determination on eyelid and positioning canthuses, realize the eyes recognition.The further study of this paper is the eyes recognition algorithm on the embedded platform based on ARM + DSP processing core Davinci system. Between the PC and embedded platforms, establish Linux cross-compiler environment; In the ARM side of the platform, the Linux-based operating system designs the image processing process of the multi-threaded architecture to image acquisition, transmission, processing and display; XDM specification package on the eyes recognition algorithms, and integrates into the multi-threaded processes, the Davinci platform Codec architecture for the processing of the algorithm in the DSP-side; Use all kinds of optimization method to enhance the efficiency of code, and the process of the optimization of the test reflect the effect.Based on the above mentioned Da Vinci platform, the eye identification in the indoor environment obtained a good test results.