原文: http://blog.163.com/ld081055@126/blog/static/11818691520105144852433/
还有这个文章也实现了:http://blog.csdn.net/lxdfigo/article/details/8279919
几个开源项目:
https://github.com/urgv/pthreads-fft2d
https://github.com/BenedictHiddleston/fft2d
https://github.com/ecordon/FFT2
这个缺少函数:trace
前面编过2D-FFT的程序,现在把2D-IFFT的程序整合到一起,便于后面做图像变换反变换使用。
FFT与IFFT有如下关系:
相应的2D-FFT与2D-IFFT的关系如下:
所以可以利用一个FFT核心函数实现2D-FFT与2D-IFFT。代码如下:
#include <stdio.h>
#include <stdlib.h>
#include <math.h>
#define intsize sizeof(int)
#define complexsize sizeof(complex)
#define PI 3.1415926
int *a,*b;
int nLen,init_nLen,mLen,init_mLen,N,M;
FILE *dataFile;
typedef struct{
float real;
float image;
}complex;
complex *A,*A_In,*W;
complex Add(complex, complex);
complex Sub(complex, complex);
complex Mul(complex, complex);
int calculate_M(int);
void reverse(int,int);
void readData();
void fft(int,int);
void Ifft();
void printResult_fft();
void printResult_Ifft();
int main()
{
int i,j;
readData();
A = (complex *)malloc(complexsize*nLen);
reverse(nLen,N);
for(i=0; i<mLen; i++)
{
for(j=0; j<nLen; j++)
{
A[j].real = A_In[i*nLen+b[j]].real;
A[j].image = A_In[i*nLen+b[j]].image;
}
fft(nLen,N);
for(j=0; j<nLen; j++)
{
A_In[i*nLen+j].real = A[j].real;
A_In[i*nLen+j].image = A[j].image;
}
}
free(a);
free(b);
free(A);
A = (complex *)malloc(complexsize*mLen);
reverse(mLen,M);
for(i=0; i<nLen; i++)
{
for(j=0; j<mLen; j++)
{
A[j].real = A_In[b[j]*nLen+i].real;
A[j].image = A_In[b[j]*nLen+i].image;
}
fft(mLen,M);
for(j=0; j<mLen; j++)
{
A_In[j*nLen+i].real = A[j].real;
A_In[j*nLen+i].image = A[j].image;
}
}
free(A);
printResult_fft();
Ifft();
printResult_Ifft();
return 0;
}
void readData()
{
int i,j;
dataFile = fopen("dataIn.txt","r");
fscanf(dataFile,"%d %d",&init_mLen,&init_nLen);
M = calculate_M(init_mLen);
N = calculate_M(init_nLen);
nLen = (int)pow(2,N);
mLen = (int)pow(2,M);
A_In = (complex *)malloc(complexsize*nLen*mLen);
for(i=0; i<init_mLen; i++)
{
for(j=0; j<init_nLen; j++)
{
fscanf(dataFile,"%f",&A_In[i*nLen+j].real);
A_In[i*nLen+j].image = 0.0;
}
}
fclose(dataFile);
for(i=0; i<mLen; i++)
{
for(j=init_nLen; j<nLen; j++)
{
A_In[i*nLen+j].real = 0.0;
A_In[i*nLen+j].image = 0.0;
}
}
for(i=init_mLen; i<mLen; i++)
{
for(j=0; j<init_nLen; j++)
{
A_In[i*nLen+j].real = 0.0;
A_In[i*nLen+j].image = 0.0;
}
}
printf("Reading initial datas:\n");
for(i=0; i<init_mLen; i++)
{
for(j=0; j<init_nLen; j++)
{
if(A_In[i*nLen+j].image < 0)
{
printf("%f%fi\t",A_In[i*nLen+j].real,A_In[i*nLen+j].image);
}
else
{
printf("%f+%fi\t",A_In[i*nLen+j].real,A_In[i*nLen+j].image);
}
}
printf("\n");
}
printf("\n");
printf("Reading formal datas:\n");
for(i=0; i<mLen; i++)
{
for(j=0; j<nLen; j++)
{
if(A_In[i*nLen+j].image < 0)
{
printf("%f%fi\t",A_In[i*nLen+j].real,A_In[i*nLen+j].image);
}
else
{
printf("%f+%fi\t",A_In[i*nLen+j].real,A_In[i*nLen+j].image);
}
}
printf("\n");
}
}
void fft(int fft_nLen, int fft_M)
{
int i;
int lev,dist,p,t;
complex B;
W = (complex *)malloc(complexsize*fft_nLen/2);
for(lev=1; lev<=fft_M; lev++)
{
dist = (int)pow(2,lev-1);
for(t=0; t<dist; t++)
{
p = t*(int)pow(2,fft_M-lev);
W[p].real = (float)cos(2*PI*p/fft_nLen);
W[p].image = (float)(-1*sin(2*PI*p/fft_nLen));
for(i=t; i<fft_nLen; i=i+(int)pow(2,lev))
{
B = Add(A[i],Mul(A[i+dist],W[p]));
A[i+dist] = Sub(A[i],Mul(A[i+dist],W[p]));
A[i].real = B.real;
A[i].image = B.image;
}
}
}
free(W);
}
void printResult_fft()
{
int i,j;
printf("Output FFT results:\n");
for(i=0; i<mLen; i++)
{
for(j=0; j<nLen; j++)
{
if(A_In[i*nLen+j].image < 0)
{
printf("%f%fi\t",A_In[i*nLen+j].real,A_In[i*nLen+j].image);
}
else
{
printf("%f+%fi\t",A_In[i*nLen+j].real,A_In[i*nLen+j].image);
}
}
printf("\n");
}
}
void printResult_Ifft()
{
int i,j;
printf("Output IFFT results:\n");
for(i=0; i<mLen; i++)
{
for(j=0; j<nLen; j++)
{
if(A_In[i*nLen+j].image < 0)
{
printf("%f%fi\t",A_In[i*nLen+j].real,A_In[i*nLen+j].image);
}
else
{
printf("%f+%fi\t",A_In[i*nLen+j].real,A_In[i*nLen+j].image);
}
}
printf("\n");
}
free(A_In);
}
int calculate_M(int len)
{
int i;
int k;
i = 0;
k = 1;
while(k < len)
{
k = k*2;
i++;
}
return i;
}
void reverse(int len, int M)
{
int i,j;
a = (int *)malloc(intsize*M);
b = (int *)malloc(intsize*len);
for(i=0; i<M; i++)
{
a[i] = 0;
}
b[0] = 0;
for(i=1; i<len; i++)
{
j = 0;
while(a[j] != 0)
{
a[j] = 0;
j++;
}
a[j] = 1;
b[i] = 0;
for(j=0; j<M; j++)
{
b[i] = b[i]+a[j]*(int)pow(2,M-1-j);
}
}
}
complex Add(complex c1, complex c2)
{
complex c;
c.real = c1.real+c2.real;
c.image = c1.image+c2.image;
return c;
}
complex Sub(complex c1, complex c2)
{
complex c;
c.real = c1.real-c2.real;
c.image = c1.image-c2.image;
return c;
}
complex Mul(complex c1, complex c2)
{
complex c;
c.real = c1.real*c2.real-c1.image*c2.image;
c.image = c1.real*c2.image+c2.real*c1.image;
return c;
}
void Ifft()
{
int i,j;
for(i=0; i<mLen; i++)
{
for(j=0; j<nLen; j++)
{
A_In[i*nLen+j].image = -A_In[i*nLen+j].image;
}
}
A = (complex *)malloc(complexsize*nLen);
reverse(nLen,N);
for(i=0; i<mLen; i++)
{
for(j=0; j<nLen; j++)
{
A[j].real = A_In[i*nLen+b[j]].real;
A[j].image = A_In[i*nLen+b[j]].image;
}
fft(nLen,N);
for(j=0; j<nLen; j++)
{
A_In[i*nLen+j].real = A[j].real/nLen;
A_In[i*nLen+j].image = A[j].image/nLen;
}
}
free(A);
free(a);
free(b);
A = (complex *)malloc(complexsize*mLen);
reverse(mLen,M);
for(i=0; i<nLen; i++)
{
for(j=0; j<mLen; j++)
{
A[j].real = A_In[b[j]*nLen+i].real;
A[j].image = A_In[b[j]*nLen+i].image;
}
fft(mLen,M);
for(j=0; j<mLen; j++)
{
A_In[j*nLen+i].real = A[j].real/mLen;
A_In[j*nLen+i].image = A[j].image/mLen;
}
}
free(A);
free(a);
free(b);
}
测试数据及结果如下:
数据输入文件data_In.txt的内容如下:
3 3
1 2 5
3 2 5
5 6 4
测试结果如下:
Reading initial datas:
1.000000+0.000000i 2.000000+0.000000i 5.000000+0.000000i
3.000000+0.000000i 2.000000+0.000000i 5.000000+0.000000i
5.000000+0.000000i 6.000000+0.000000i 4.000000+0.000000i
Reading formal datas:
1.000000+0.000000i 2.000000+0.000000i 5.000000+0.000000i 0.000000+0.000000i
3.000000+0.000000i 2.000000+0.000000i 5.000000+0.000000i 0.000000+0.000000i
5.000000+0.000000i 6.000000+0.000000i 4.000000+0.000000i 0.000000+0.000000i
0.000000+0.000000i 0.000000+0.000000i 0.000000+0.000000i 0.000000+0.000000i
Output FFT results:
33.000000+0.000000i -5.000000-10.000000i 13.000000+0.000000i -5.000000+10.000000i
-7.000000-10.000000i -7.000000+6.000000i 1.000000-6.000000i -3.000000-2.000000i
13.000000+0.000000i -1.000000-6.000000i 1.000000+0.000000i -1.000000+6.000000i
-7.000000+10.000000i -3.000000+2.000000i 1.000000+6.000000i -7.000000-6.000000i
Output IFFT results:
1.000000+0.000000i 2.000000+0.000000i 5.000000+0.000000i 0.000000-0.000000i
3.000000+0.000000i 2.000000+0.000000i 5.000000+0.000000i 0.000000-0.000000i
5.000000+0.000000i 6.000000+0.000000i 4.000000+0.000000i -0.000000-0.000000i
0.000000-0.000000i 0.000000+0.000000i 0.000000-0.000000i -0.000000+0.000000i
Press any key to continue