基于Gabor滤波与灰度梯度共生矩阵的鸽眼虹膜识别

Pigeon Iris Recognition Method Based on Multi-Channel Gabor Filtering and GGCM

禽畜标识技术作为当前研究热点难点受到广泛关注,以虹膜识别为代表的生物特征标识技术成为一种趋势。针对传统方法提取全局特征时对局部纹理特征不敏感的问题,文中提出了一种基于多通道Gabor滤波和灰度梯度共生矩阵(GGCM)的鸽眼虹膜识别方法。使用不同方向和尺度的Gabor滤波器组对预处理后的虹膜图像进行滤波,由全局滤波图像构建灰度梯度共生矩阵,提取其特征值生成虹膜特征向量,识别过程采用K-NN分类算法。试验结果表明:该方法具有识别精度高,等错率低等优点,在候鸟监测、家禽管理和肉食品追溯方面具有一定的实际应用价值。

Livestock identification technology attracted widespread attention as the current research hotspot and difficult. As the representative of biometric identification, the animal iris recognition tends to become a new identification technology. In order to solve the problem that local texture is ignored when it comes to the traditional method to extract the global features, a pigeon iris recognition method based on multi-channel Gabor filtering and GGCM was proposed. First of all, the feature image was generated through different scales and directions of Gabor filter; Then it was built GGCM by the global filtered image and extract feature value to generate the iris feature vector. The recognition process used K- NN classifi- cation algorithm. Experimental results showed that the method had high identification accuracy and low error rate, and had certain practical application value in the aspect of migratory bird monitoring, poultry and meat dish traceability management.