Research On Technology Of Cattle's Iris Recognition

The demand of information security is increasing. The traditional identification methods can not meet the requirements of social development because of their instability and the inherent characteristics themselves. That's the reason that the biometrics emerged. So far, the biometrics such as face, palm print, fingerprint and voice has been used in various fields. The most important one is iris recognition. It has lots of advantages like uniqueness, stability, non-invasive and so on. Compared to other biometrics, it contains more unique features and higher accuracy. According to the statistics, iris recognition has the lowest error rate among all biometric identification.However, currently the research of iris recognition technology focuses on human iris, only few for that of animals. With the prevalence of mad cow disease in Europe, the issue of intensifying the food safety has been paid much more attention by many countries. How to increase the accuracy of animal meat product traceability and reduce the cost of the products has become an urgent issue to be resolved. As one of the novel solution to the issue, animal iris recognition tends to become a new animal identification technology, which has potential applications to control the safety of the meat products and has great advantages compared to the traditional methods.This paper discusses the prospects and significance of cattle iris recognition firstly, which includes the theory of iris recognition and the technical difficulties in cattle iris recognition. Then, we introduce the key technology in all stages of cattle iris recognition such as cattle iris image acquisition and pre-processing, cattle iris segmentation, cattle iris normalization and enhancement, cattle iris feature extraction and matching.In these stages, iris image acquisition and pre-processing are the basis of cattle iris recognition. The quality of iris image acquisition will affect the accuracy of the recognition results. The purpose of cattle iris segmentation is to extract the cattle iris from the cattle image. In this paper, some improvements have been presented. An approach based on Least Square Principle has been proposed for cattle iris segmentation. The distance between the edge points and the fitted curve is calculated to decide the best one in the cattle iris image. The comparison of fitting precision and speed between the proposed approach and traditional methods is showed in the experiments. In cattle iris normalization and enhancement stage, a geometric method is used to make the cattle iris normalize to a rectangular block to achieve shift and scale invariance. Finally, the Gabor filter is used for feature extraction and weighted Euclidean distance or improved Euclidean distance can be used for matching.Experimental results indicate the efficiency and quick speed of the proposed approach compared to the traditional methods.