本文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-geometric model of this matching method for the study and submit a complete set of facial positioning algorithm. In preprocessing, using a special facial features enhanced facial skin and the contrast between the methodology and obtain local threshold value of the two methods, improved the pretreatment results. Image segmentation, the realization of the classical division algorithm, and the use of group division algorithm improved results. Face in the match, based on the realization of the eyes and mouth of the geometric model matching and evaluation function of the tectonic study