clear all;
clc;
close all;
%x=0.01:0.01:10;
%y1=linspace(0,0,120);%120
%y2=0:-0.004:-0.5;%126 位
%y3=linspace(-0.5,-0.5,128);%128 位
%y4=-0.5:0.004:0.5;%251 位
%y5=linspace(0.5,0.5,128);%128 位
%y6=0.5:-0.004:0;%126 位
%y7=linspace(0,0,121);%121
%y=[y1,y2,y3,y4,y5,y6,y7];
x=0.01:0.01:30;
y1=linspace(0,0,298);%298
y2=0:-0.0065:-0.65;%100 位
y3=linspace(-0.65,-0.65,300);%300 位
y4=-0.65:0.0065:0;%100 位
y5=linspace(0,0,1400);%1400 位
y6=0:0.0065:0.65;%100 位
y7=linspace(0.65,0.65,300);%300
y8=0.65:-0.0065:0;%100 位
y9=linspace(0,0,298);%298
y=[y1,y2,y3,y4,y5,y6,y7,y8,y9];
figure(1)
subplot(2,1,1)
plot(x,y)
L=length(y); %Calculate the length of an audio signal
noise=0.02*randn(1,L); %Generate a random noise signal of equal length (the size of the noise here
depends on the amplitude multiple of the random function）
subplot(2,1,2)
plot(x,noise)
y_z=y+noise; %Superimpose two signals into a new signal - add noise processing
figure(2)
subplot(2,1,1)
plot(x,y_z)
%N = normalize(y_z)
%Nr = normalize(y_z,'range')
%subplot(2,1,2)
%plot(Nr)
fc=var(y_z);%To find the variance of the entire signal, it is not used temporarily, it may be used when
determining the threshold
j=0;
for i=1:L;
if (y_z(i)<-0.0468);
j=j+1;
m(j)=y_z(i);
end
end
L1=length(m) ;
m1= resample(m, L, L1);%Scale normalization processing, resample is generated by resampling m by
L/L1 times m1
subplot(2,2,3)
plot(m)
subplot(2,2,4)
plot(m1)
N=7; %Representatives are divided into five layers
f=zeros(7,L1); %fRepresents the low frequency band and is divided into five layers zeros represents
the generation of an all-zero matrix, 5 rows and 1000 columns
h=zeros(7,L1); %hOn behalf of the high frequency band, it is divided into five layers
[c,l]=wavedec(m,N,'haar');%Commonly used wavelets haar、db1 4 8 15、coif1 2 3、sym1 2
%Perform one-dimensional N-scale decomposition on the signal with y3, and return the
wavelet coefficients of each layer, haar stands for Haar wavelet, which is the selected wavelet basis
function
%Return value on the left: C is each coefficient after wavelet decomposition, L is the number
of corresponding wavelet coefficients;
for i=1:7 %i take 1 to 6 respectively and do the following loop operation
f(i,:)=wrcoef('a',c,l,'haar',i);%all elements of the ith row of matrix f
h(i,:)=wrcoef('d',c,l,'haar',i);%wrcoef is a one-dimensional multi-level decomposition and
reconstruction signal function, which uses the c and l values just decomposed to reconstruct the
original signal, that is, the return value on the left；
% 'a' or 'd' represents "low frequency approximation" or "high frequency detail", C and L are the
values just decomposed, 'haar' means consistent with the wavelet basis function during
decomposition, and the last number i is where the part is located series;
end
figure(3) %Create figure1；
subplot(3,2,1); %Divide the current figure into a 3-by-2 grid and create the axes at position 1.
Number subplot positions by line number。
%The first subplot is the first column of the first row, the second subplot is the second
column of the first row, and so on.
%If axes already exist at the specified location, this command makes the axes the current
axes。
plot(x,y_z,'r'); %In the subgraph window, draw a function image with x as the abscissa and y3 as the
ordinate, and 'r' represents the use of a red line
title('原始信号');%The subplot is titled 'Original Signal’
xlabel('时间 s'); %When the x-axis is marked '
