knn

reconnaissance de parole par des HMM

说明：  先对图像均衡化 增强对比度 加噪声 然后分别进行空域和频域滤波(First, the contrast enhanced image equalization plus noise and then separate airspace and frequency domain filtering)

DFT进行频谱分析时的三种现象 1、混叠：对连续信号采样，要求连续信号是带限的，采样频率要足够高。Fs应满足Nyquist采样定理才不产生混叠。 采样前加低通滤波器防混叠 2、频谱泄漏：DFT对时域信号进行了截断（相当于采样信号乘了一个窗函数），时域的乘积相当于频域卷积，改变了原信号频谱。 改用其他窗函数（非矩形）改善泄漏。 3、栅栏现象：DFT 只给出了频谱在采样点上的取值，采样点间的频谱内容丢失。(DFT spectrum analysis of a three phenomena, aliasing : continuous signal sampling, requiring continuous signal is band-limited, the sampling frequency should be high enough. Fs meet the Nyquist Sampling Theorem would not have aliasing. Sampling former low-pass filter plus anti-aliasing 2, spectrum leakage : DFT right time domain signal a disconnect (equivalent to a sampling signal by a window function). time domain is equivalent to the product of frequency-domain convolution, changes in the spectrum of the original signal. Window functions to switch to other (non-rectangular) to improve leakage. 3, fenced phenomenon : DFT spectrum is given only at the sampling point value, sampling points of the spectrum as lost.)

小波包分析提取振动信号中的特征频率，真的是一个好程序，窗函数法设计一个数字带通FIR滤波器。( Wavelet packet analysis to extract vibration signal characteristic frequency, Really is a good program, A window function design FIR digital band-pass filter.)

对于连续非线性系统，可以使用灰色预测PID控制(For a continuous nonlinear system, you can use gray prediction PID control)

说明：  LMS自适应算法的matlab仿真源程序(LMS adaptive algorithm matlab simulation source)

正交扩频码ooc的matlabb代码。五个用户情况 (Orthogonal spread spectrum the code ooc matlabb code. Five users)

基于matlab，利用背景减除法实现运动目标的跟踪。本代码所用的函数均适用于新版本的matlab，不需要额外修改。在MATLAB2014a编译通过。(Based on matlab, for tracking moving objects using background subtraction method. The code used functions are applied to the new version of matlab, no additional modifications. In MATLAB2014a compile.)

网络分析法(The Analytic Network Process，ANP)的Matlab源代码(Network analysis (The Analytic Network Process, ANP) of the Matlab source code)

计算矩阵序列的低秩近似，来至于论文一般的矩阵序列低秩近似。用matlab实现，能够有非常高的压缩率，而且能够重建原图像(We consider the problem of computing low rank approximations of matrices.)