MATLAB_GUI设计-专题讲稿-王正盛

主成分分析 （ Principal Component Analysis ， PCA ）或者主元分析。是一种掌握事物主要矛盾的统计分析方法，它可以从多元事物中解析出主要影响因素，揭示事物的本质，简化复杂的问题。计算主成分的目的是将高维数据投影到较低维空间。给定 n 个变量的 m 个观察值，形成一个 n ′ m 的数据矩阵， n 通常比较大。对于一个由多个变量描述的复杂事物，人们难以认识，那么是否可以抓住事物主要方面进行重点分析呢？如果事物的主要方面刚好体现在几个主要变量上，我们只需要将这几个变量分离出来，进行详细分析。但是，在一般情况下，并不能直接找出这样的关键变量。这时我们可以用原有变量的线性组合来表示事物的主要方面， PCA 就是这样一种分析方法。(Principal component analysis (Principal Component Analysis, PCA) or PCA. Is a statistical method to grasp the principal contradiction of things, it can be resolved diverse things out the main factors, revealing the essence of things, simplifying complex problems. The purpose of calculating the main component of high-dimensional data is projected to a lower dimensional space. Given n variables of m observations, forming an n ' m of the data matrix, n is usually large. For a complex matters described by several variables, it is difficult to know, so if you can grab something to focus on key aspects of analysis? If the main aspects of things just reflected on several key variables, we only need to separate out these few variables, for detailed analysis. However, in general, does not directly identify this critical variables. Then we can represent the major aspects of things with a linear combination of the original variables, PCA is one such analysis.)

很好用的svm工具箱，dagsvm对于图像分类效果非常好，分类精度很高，代码清洗简单。(Good use of SVM Toolbox, dagsvm for image classification effect is very good, very high classification accuracy, code cleaning easy.)

进行基于绝缘子边缘检测与提取,能识别绝缘子裂纹的故障,通用性蛮好(Insulator-based edge detection and extraction, to identify cracks insulator failure, versatility fine)

一维信号的形态学滤波实现，包括开，闭，腐蚀，膨胀功能(morphology filter for 1-D signal,include open,close,Corrosion and dilate)

低秩学习，rpca，适用于图像去噪，视频跟踪。(low rank learning,rpca,image denoise.Image tracking)

视频中静态目标背景下运动的物体的前景提取算法的实现。(The foreground extraction algorithm for moving objects in static background is implemented in video.)

　找出图象的最佳阈值，并显示它的二值图．(find the best image threshold, and it shows the value of Figure 2.)

利用固定目标和运动目标之间速度的差别所引起的多普勒频移不同根据在动目标 显示和动目标检测中使用的各种滤波器去除固定杂波提取动目标的回波能够大大 改善在杂波背景下检测运动目标的能力仿真结果显示出很好的效果也给出了改善 因子与积累脉冲数间关系 (radar moving target simulation)

包括影像预处理，叶绿素反演，水污染监测等功能(Including image preprocessing, water pollution monitoring and other functions)

非常好的人脸识 别的代码，有详细的注释，还有图片库，好东西就要大家分享！(Face recognition is very good code, there are detailed notes, as well as image libraries, will be a good thing to share!)