NBI

：给出了一种提取特征点的新方法和合理的特征不变量。该方法根据三角不等式原理利用第三边与另 外两边之和的比值大小来确定角点，以特征点和它相邻的两个特征点所形成的两条直线的夹角正切值以及三角形 第三边与另两边和的比值作为特征不变量。这种特征提取的方法简单，并且由于扩大了模板取值范围，使识别效果受特征点漏检的影响较小。实验表明，该方法不仅简便，而且效果良好。(: In this paper, a new feature point extraction method and the characteristics of rational invariants. The method is based on the principle of triangle inequality with the use of the third side and both sides of the ratio of the size to determine the Corner to feature point and its two adjacent feature points formed by the angle between two tangent linear and the third triangle on both sides and edges with another feature of the ratio as a variable. This feature extraction method is simple, and as a result of the expansion of the scope of the template values to identify the effect of feature points by the impact of smaller undetected. Experiments show that the method is not only simple, but effective.)

EMD 和 Hilbert-Huang 变换 matlab工具箱(EMD and HilbertHuang transform matlab toolbox)

说明：  本程序包括：论文SVM 用于基于块划分特征提取的图像分类，和相应的matlab实现其中图像划分以及特征提取、聚类均利用matlab6.5完成。 (The procedures include: paper by SVM for feature extraction based on block classification, and the corresponding realization of one image into matlab, and feature extraction, clustering were done using matlab6.5.)

matlab中计算并且标出图片中的地平线(matlab to calculate and mark the picture horizon)

channel flow computation

planovanie trajektorie mobilneho robota

It uses kalmanf(s) function to calculate GPS error variance.

MIMO Source Number detection for multiple inputs and multiple outputs

车辆系统动力学主动悬架线性最优控制（LQG）程序，在simulink环境下仿真计算出路面输入下的最优主动悬架性能，并于传动被动悬架作出对比(Vehicle System Dynamics active suspension linear optimal control (LQG) program, simulink simulation environment optimal active suspension performance under the road input drive passive suspension)

1 通过实验，掌握多元正态分布的最大似然估计； 2 掌握多元正态分布下的最小错误率的贝叶斯分类； 3 对其他的参数估计有更深的认识。 (1 experiment, master multivariate normal distribution maximum likelihood estimation 2 multivariate normal distribution under the minimum control error rate Bayesian classifier 3 on the other have a better understanding of parameter estimation.)