Research On Ear Recognition Algorithm And Application System Establishing

The ear recognition is a biometrics identification method which be concerned by public from2000, due to the ear has good stability in its growth process, strong anti-interference in the identification process and unique biological characterlistic, it can achieve effective identification. In addition, the ear recognition technology is nonintrusive identification which has low demand in acquisition of equipment and the acquisition conditions can quickly identify.It has high practical application value.The ear recognition algorithm is researched in this paper, and the ear image preprocessing algorithm and feature recognition algorithm is analyzed. Sampling, pretreatment, object extraction, database construction and identification complete algorithms are designed and simulated, the effectiveness of the algorithm is proved by simulation results. Build the hardware system and program the algorithm, the ear recognition system is designed. Test results of the system show that:The system is achieved to identify the target identity by the identification of the ear. The main content is as follows:1. Completion an algorithem for ear image preprocessing. Build an ear color model, and the ear region are processed by filter and morphology denoising. Then the processed image can be accurate positioned by edge detection, and the target image is normalized.2. Completion of the ear image feature extraction. Analysis of the practical application of the ear to identify circumstances under controlled conditions, selecting the PCA (principal component analysis) to extract the features of the ear image.3. Completion of the ear image database construction and matching. According to the ear database construction standards, established experimental ear image database. The ear image database is matched by using the K-nearest neighbor classification method based on Euclidean distance; the matching results are counted and analyzed.4. Completion of the ear recognition system. Build the hardware system and program the algorithm. The system can cellect ear images under real-time and complete identity recognition technology based on ear recognition technology.Using the system to recognize the ear image in an experimental environment, the positive ear recognition rate reaches82.3%, the accumulation recognition rate reaches97.8%. The system can achieve the requirement of recognition and identification.