联合灰度与梯度性质的掌纹主线提取方法

Principal lines extraction method of palmprint based on gray and gradient properties

掌纹主线的提取可为掌纹识别、掌纹诊断专家系统等掌纹处理技术提供重要的数据。为了让提取的掌纹主线结果形态自然,位置正确,提出利用Laplacian算子对掌纹灰度图像进行了对比度增强,采用局部灰度平方差来分割出掌纹主线,通过Sobel算子与范围模板去除噪声,对获得的特征图像采用数学形态学操作并细化,从而获取简化的主线特征。对不同形态的掌纹进行了算法验证,结果表明该方法保留了掌纹主线的细节信息,能满足后续处理技术的数据要求,并且算法简单、实时性强。

The principal lines of palmprint extraction can provide important information for palmprint identification and palmprint diagnosis expert system. In order to extract palmprint with natural shapes and real position, Laplacian operator was used to sharpen contrast of gray image, then gray square difference algorithm was applied to split the palmprint features of the image. The gray noise was removed by using the Sobel operator and the range template. Mathematical morphology operation and thining algorithm were used to obtain the simplified main line features. In the last, the different morphology of palmprint was verified by our algorithm, and the experimental results prove its reliability for saving detailed characteristics. The algorithm is simple and fast, and it can satisfy the requirements of real-time processing and follow-up work.