基于几何匹配和分合算法的生物特征识别技术在公安视频监控中人脸识别的应用研究
发布时间:2018-04-25 19:20
本文选题:人脸识别 + 几何匹配 ; 参考:《中南大学》2010年硕士论文
【摘要】: 本文研究基于几何匹配和分合算法的生物特征识别技术,及其在公安视频监控中人脸识别的应用。本文首先对几何模型匹配方法进行了研究,提出了一套完整的人脸定位算法。在预处理部分,采用了特别的增强人脸特征与脸部皮肤之间对比度的方法及局域取阈值二值化方法,改进了预处理的效果。在图像分割部分,实现了经典的分合算法,并且使用成组算法改进了分合的效果。在人脸匹配部分,实现了基于眼睛和嘴的几何模型匹配,并对评价函数的构造进行了研究。 在Microsoft Windows XP平台上Microsoft Visual C++开发环境下,采用上述算法编制了人脸定位的软件。实验证明,该软件对于尺寸足够大的清晰的正面单个/多个人脸情形能够正确检测定位。 最后,本文探讨了几何匹配算法和分合算法的优缺点,及人脸识别在公安工作中的应用,提出了进一步工作的方向。
[Abstract]:This paper studies the biometric recognition technology based on geometric matching and partitioning algorithm, and its application in face recognition in police video surveillance. Firstly, the geometric model matching method is studied, and a complete face location algorithm is proposed. In the preprocessing part, a special method of enhancing the contrast between facial features and facial skin and a local threshold binarization method are used to improve the effect of pretreatment. In the part of image segmentation, the classical partitioning algorithm is implemented, and the effect of grouping algorithm is improved. In the part of face matching, the geometric model matching based on eyes and mouth is realized, and the construction of evaluation function is studied. In the Microsoft Visual C development environment of Microsoft Windows XP, the software of face location is programmed by the above algorithm. The experimental results show that the software can detect the location correctly for the large enough size of the single / multiple faces. Finally, this paper discusses the advantages and disadvantages of geometric matching algorithm and partitioning algorithm, and the application of face recognition in public security work, and puts forward the direction of further work.
【学位授予单位】:中南大学
【学位级别】:硕士
【学位授予年份】:2010
【分类号】:TP391.41
【引证文献】
相关期刊论文 前1条
1 梁永霖;;基于OpenCV的人脸图像预处理技术研究[J];韶关学院学报;2012年08期
,本文编号:1802616
本文链接:https://www.wllwen.com/falvlunwen/fanzuizhian/1802616.html