在现代化技术高速发展的今天，人们的安全意识逐渐增强，公共安全也成为人们关注的焦点。有效的身份识别技术也成为了共同关注的热点。与传统的身份识别技术相比步态识别具有非侵犯性，难以隐藏等特点。所以身份识别具有重要的研究意义和广泛的应用前景。本文以提高步态识别精度为目标，以提取步态特征为重心，提出了主成分分析(PCA)与线性判别分析(LDA)融合算法和二维主成分分析(2DPCA)与二维线性判别分析(2DLDA)融合算法，来提取步态特征，主要做了以下工作： 首先，步态图像预处理及能量图(GEI)计算。本文选用了背景减除法分割运动目标且将分割后的图像进行二值化处理和形态学操作，获得比较完整的二值步态图像序列，根据每帧图像中运动目标的宽度变化来计算步态周期。将一个步态周期内的所有图像序列累加求和得到步态能量图，作为一个人的步态特征。 其次，步态特征提取及实验结果分析。为了能够有效的进行人体步态识别研究，提出一种基于PCA与LDA的步态识别方法，使用这个方法，在已得到GEI的基础上，提取步态特征，为了解决线性判别“小样本”问题，又提出了2DPCA与2DLDA的融合算法。然后使用K邻近分类器对待测试图像进行分类识别。 最后，搭建步态识别系统。该系统将图像预处理、步态周期检测、步态能量图计算、以及步态特征的提取和分类识别中所用到的算法，组合成一个完整的步态识别系统。在该系统中可以查看每一个功能的效果图。 实验结果表明，在中科院提供的CASIA DataBaseB中，采用本文提出的两种融合算法能够有效的提高步态识别率，且在不同角度、不同状态下都能得到较高的识别率。

Owing to the rapid development of modern technology, people's safety awareness is gradually enhanced,the research area of Public security has become a hot spot. Effective identification technology has become a hot topic of common concern. Having compared with the traditional identification ways, gait identification technology to identify non-invasive, and it is difficult to hide, so it has important significance and broad application prospects in the physiological characteristics detection and identification. In this paper, focus on the gait feature extraction, Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) algorithm fusion and fusion algorithm two Dimensional Principal Component Analysis (2DPCA) and two Dimensional Linear Discriminant Analysis is proposed (2DLDA), in order to improve gait recognition accuracy as the goal to extract gait characteristics, the main the following work： Firstly, the gait image preprocessing and Energy calculations. The background subtraction method is used to get the object segmentation, segmentation out binary image processing and morphological operations to obtain a more complete and clear binary gait image sequences, the width change of each frame image of the moving object is used to calculate the motion cycle, all the pixels image sequence a gait cycle has will be sumed to get the gait energy image with all gait characteristics of a person. Secondly, the gait feature extraction and analysis of experimental results. To be able to effectively carry out human gait recognition, gait recognition method is based on proposed gait energy image and PCA and LDA. Will advance to use the gait energy image we can get earlier, the use of principal component analysis and linear discriminant analysis fusion algorithm to extract gait characteristics, in order to solve the problem of linear discriminant "small sample". 2DPCA of 2DLDA fusion algorithm is proposed. Then,use k-Nearest Neighbor algorithm (KNN) to get the image classification. Finally, build gait recognition system. The system will be mentioned in this the image processing, the gait cycle analysis, gait energy image computing, and gait feature extraction and classification, these will be combined together to get a complete gait recognition system. In this system, you can see the results after each function processed. Experimental results show that with the CASIA database B provided by the Chinese Academy of Sciences, using two fusion proposed method can effectively improve the gait recognition rate, in different angles, different states, high recognition rate can be achieved