了适应跟踪过程中目标光照条件的变化，并对目标特征进行在线更新，提出一种将局部二元模式(LBP)

特征与图像灰度信息相融合，同时结合增量线性判别分析对目标进行跟踪的算法．跟踪开始前，为了获得比较准确的目标描述，使用混合高斯模型和期望最大化算法对目标进行分割；跟踪过程中，通过蒙特卡罗方法对目标区域和背景区域进行采样，并更新特征空间参数．得到目标和背景的最优分类面；最后使用粒子滤波器结合最优分类面对目标状态进行预测．通过光照变化的仿真视频和自然场景视频的跟踪实验，验证了文中算法的有效性．-Tracking process to adapt to changes in the target lighting conditions, and the target feature for online updates, proposes a local binary pattern (LBP) features and image intensity information integration, combined with incremental linear discriminant analysis for target tracking algorithms. Track begins, in order to obtain a more accurate descr iption of the objectives, the use of Gaussian mixture models and expectation maximization algorithm for target segmentation　tracking process, through the Monte Carlo method of the target area and the background area sampled and updated feature space parameters. Get the optimal target and background classification surface　finally Using Particle Filter optimal classification predict the state of the face of goal. By varying illumination simulation video and natural scenes video tracking experiment to verify the effectiveness of the proposed algorithm.