《数字图像处理 Digital Image Processing》课件 Frequency Domain Enhancement.ppt

Result of GHPF filtering Note that the smaller objects and thin bars is cleaner with the Gaussian filter. Lecture 9: Chapter 4 Frequency Domain Enhancement fig4_26_demo.m The Laplacian in the frequency domain From mathematical theorem, we know: For a two variable function, it follows that: Lecture 9: Chapter 4 Frequency Domain Enhancement The expression inside the brackets happen to be the Laplacian of f(x,y). Thus, we have the important result: The above result means that the Laplacian can be implemented in the frequency domain by using the filter: H(u, v) = －(u2 + v2). The Laplacian derivation con’t Lecture 9: Chapter 4 Frequency Domain Enhancement The Laplacian continues As before, the center of the Fourier transform is shifted, the center of the filter function also needs to be shifted: The Laplacian-filtered image in the spatial domain is obtained by computing the inverse Fourier transform of H(u,v)F(u,v): Lecture 9: Chapter 4 Frequency Domain Enhancement Conversely, computing the Laplacian in the spacial domain and then computing the Fourier transform of the result is equivalent to multiplying F(u,v) by H(u,v). We express this dual relationship in the familiar Fourier-transform-pair notation The Laplacian continues Lecture 9: Chapter 4 Frequency Domain Enhancement Demo: freq_laplace.m Pictorial representation of the Laplacian The function is centered at (M/2, N/2), and its value at the top of the dome is zero. All other values are negative. Note the gray-level profile in fig. e), This corresponds to the spatial mask form shown in fig. f). Lecture 9: Chapter 4 Frequency Domain Enhancement Laplsfq.m The Laplacian application in the frequency domain Laplacian-filtered images contain both positive and negative values of comparable magnitudes, scaling is necessary. Lecture 9: Chapter 4 Frequency Domain Enhancement Frequency Laplacian demonstration A=imread(Fig3.40(a).jpg); imshow(A); A=im2double(A); A1=fft2(A); A1=fftshift(A1); [m,n]=size(A); H=fqlaplas(m,n); B