coe_gyT

研一上学期，学校开了数值分析这门课，而且要求用matlab进行上机试验，经过一段时间的学习，对matlab有了最初步的了解，结合所学课程，废了好大的劲才编了几个小程序，希望对后来学习的人有点帮助。(Kenichi last semester, the school opened the numerical analysis of the course, but also carried out on machines with matlab test, after a period of learning, of matlab with the most basic understanding, combined with the course, the repeal of several large series劲才a few small programs, and they hope to learn a little bit later to help.)

说明：  MATLAB使用百例，1-32是：图形应用篇；33-66是：界面设计篇；67-84是：图形处理篇； 85-100是：数值分析篇(MATLAB using the 100 cases, 1-32 is: graphics applications, articles 33-66 is: interface design articles 67-84 is: graphics processing articles 85-100 are: Numerical Analysis Section)

在排序学习算法经常用到的评价指标 NDCG，计算检索的效果好坏(NDCG can be used to evaluate the ranking with multilevel relevance such as irrelevance, half relevance, and relevance. )

这里面的matlab带啊能够实现图像的二值化编程，对图像处理的朋友有帮助(This is inside the matlab image can be achieved with ah binary programming, image processing helpful friend)

雷达系统设计MATLAB仿真一书里的第四张所有的程序，对于学习非常有用。(Radar system design MATLAB simulation fourth book program, very useful for learning.)

Good research paper for cooperative spectrum sensing for cognitive radio

香农—费诺编码的MATLAB实现，程序已在，可直接运行(Shannon- Fenno encoded MATLAB realization of the process had been in the direct operation)

说明：  建立关联程序的例子。本程序演示把 .bar 文件与 NotePad 建立关联(establish procedures related examples. The demonstration put procedures. Bar documents associated with the establishment of NotePad)

基于小波变换的分开—合并图像分割matlab．采用多尺度小波变换系数作为四分树结构的分开一合并法图像分割的一致性度量 从而在大的图像块中。。去除噪声的影响，在小的图像块中，以小波变换的局部极大值精确定位边缘，根据边缘信息进行分开一合并法图像分割 ．实验表明．算法得到满意的结果 ． (Separately based on wavelet transform- merge image segmentation Matlab. Multi-scale wavelet transform coefficients as the quadtree structure separate consolidation method of image segmentation consistency measure large image block. . To remove the effects of noise, in small image blocks, the local maxima of the wavelet transform precise positioning edge, to separate a combined method for image segmentation according to the edge information. Experiments show. Algorithm to obtain satisfactory results.)

：滤波 fs=400 采样频率,必须大于两倍基带信号最高频率 ts=1/fs 采样时间间隔 T=5 时间窗大小 t=0:ts:T (hf1=zeros(1,N) hf1(f<=B1&f>=-B1)=1 y1=hf1.*fftshift(y) figure subplot(211) plot(f,abs(fftshift(y))/N,f,hf1, r-- ) axis([-fs/2,fs/2,0,1.2]) title( 原信号频谱 ) xlabel( f /Hz ) subplot(212) plot(f,abs(y1)/N) title( 低通滤波，截止频率90Hz ) xlabel( f /Hz ))