| 摘要 |
<p>PURPOSE:To improve the computing accuracy of the 2-dimensional fast Fourier transformation (FFT) by shifting up a remaining part excluding a DC component by (n) bits after the 1-dimensional FFT performed in the 1st direction to change the position of a fixed decimal point and then carrying out the 1-dimensional FFT in the 2nd direction. CONSTITUTION:A 1-dimensional FFT operation is carried out in the 1st direction (x) to a 2-dimensional picture by a fixed decimal point system. The another 1-dimensional FFT operation is carried out in the 2nd direction (y) orthogonal to the direction (x) by the fixed decimal point system to the result of the first 1-dimensional FFT operation carried out after a part excluding an obtained DC component is shifted up by (n) bits. Thus the 2-dimensional FFT result is obtained. Then a 1-dimensional inverse FFT (IFFT) operation is carried out in the 2nd direction by the fixed decimal point system to the 2-dimensional FFT result, and a part excluding an obtained DC component is shifted down by (n) bits. Then a 1-dimensional IFFT operation is carried out in the direction (x) to the result of the 1-dimensional IFF operation. Thus the 2-dimensional IFFT result is obtained. Then it is possible to perform an optical FFT analysis of a 2-dimensional picture with high accuracy via a 1-dimensional FFT operation using a fixed decimal point system.</p> |
