武汉理工大学数字图像处理英文课件 Chapter1（3）.ppt

Coordinate Images Gaus image Image of sinc(x)2sinc(y)2 Logarithmic Enchancement Matlab function logim Special Functions Special Function Construction Somb function log image display Expanded Domain * Coordinate images have z-values proportional to x and y coordinate values. Suppose we want to define the domain of the image as a rectangular area -2.5 = x = 2.5 and -2.5 = y = 2.5: ???????????????????????????????? ???????????????????????????????? x.jpg y.jpg x = linspace(-2.5,2.5,256); y = linspace(2.5,-2.5,256); x = ones(256,1)*x; y = y * ones(1,256); imshow((x+2.5)/5); imwrite((x+2.5)/5,‘x.jpg); imshow((y+2.5)/5); imwrite((y+2.5)/5,‘y.jpg); This example generates an image of a two-dimensional gaussian function: rsq = x.*x + y.*y; imshow(sqrt(rsq)/5); imwrite(sqrt(rsq)/5,‘r.jpg); z = exp(-rsq);imshow(z); imwrite(z,‘gauss.jpg); z = sinc(x).*sinc(y); z = z.^2; imshow(z); imwrite(z,‘sincxy.jpg); The sinc function has a ring structure which can be seen more clearly by taking the logarithm of the image. The method is v = 1.0+log10(v)/decades where v is a pixel value and decades is the number of decades to stretch. decades = 4.0;v = z; t = 10^-decades; v = z.*(z=t) + (zt)*t;v = 1 + log10(v)/decades; imshow(v); imwrite(v,‘log_sincxy.jpg); function out = logim(image,decades,vref) %LOGIM Image with logarithmic compression % out = logim(image,D); % produces an output image compressed to D decades. % D must be a non-negative integer (default 4). % If D is zero, no compression is performed, but the image % is scaled to a maximum value of 1.0. % % out = logim(image,D,vref); % divides the input image by vref rather than scaling it to 1.0 % % If no output assignment is made, the output is displayed. % Author: John S. Loomis (8 March 2000) if nargin 3 vref = max(image(:)); end if nargin 2 decades = 4; end image = image / vref; if decades = 0 % truncate negative values v = image.*(image0); else t = 10^-decades; v = image.*(image=