摘要
目前,基于智能生物特征技术的指纹识别系统逐渐受到广泛的研究,已应用在银行金融、案件侦缉和安检门禁等场合,但时而未能做到100%快速识别指纹。提出一种快速处理并识别指纹的方法,先对指纹进行背景分割、均衡处理、高斯收敛、图像平滑、图像增强、图像二值化和图像细化等一系列的图像预处理,去除图像噪声和部分伪特征点,然后提取其中断点、端点和分叉点等局部细节特征点信息并使用离线训练好的BP人工神经网络进行指纹的识别。采用微软Visual Studio10的MFC模块程序对指纹识别的整个过程进行验证,效果良好,具有一定的应用研究价值。
Nowadays, the intelligent fingerprint identification based on biometrics technology is also a widely attention gradually, and has applied in the commercial banking and finance, criminal investigation department, public security access. But it sometimes fails to achieve a 100% recognition rate. Adopts a series of pre-processing methods that include background segmentation, equalization, gauss convergence, image thinning, etc., and also adopts image de-noising, image pseudo feature points eliminating, and then uses feature extraction and BP neural network to match and recognize the fingerprint feature points. The paper also verifies the whole procedure of the fingerprint identification by Microsoft Visual Studio 10, gets a good effect and is worth studying for novices.
基金
惠州市科技计划项目(No.2014-01)
关键词
指纹识别
神经网络
特征提取
Fingerprint Identification
Neural Network
Feature Extraction
