gscale

一、处理可变数量的输入和输出

利用nargin和nargout进行检测

T = testhv(4, 5)

该函数体中使用nargin返回2，使用nargout返回1.

函数nargchk可用于一个M函数体中，以检测传递函数的参量数目是否正确。

msg = nargchk(low, high, number)

参量number介于low与high之间返回空矩阵，否则返回错误信息。通常可这样使用：

function G = testhv2(x, y, z) ... error(nargchk(2, 3, nargin)); ...

键入仅有一个输入变量的语句

testhv2(6);

将产生错误消息

Not enough input arguments

同时执行终止。

通常，写出具有可变数目的输入变量和输出变量的函数时十分有用的。这里，我们使用变量varargin和变量varargout。varargin和varargout鼻血使用小写形式。例如，

function [m, n] = testhv3(varargin)

将输入的变量数读取到函数testhv3中，而

function [varargout] = testhv4(m, n, p)

则通过函数testhv4返回输出的变量数。若函数testhv3有一个固定的输入变量x，后跟输入变量的可变数目，则调用

function [m, n] = testhv3(x, varargin)

函数时，会导致varargin由用户提供的第二个输入变量开始运行。varargout的情形与此类似。一个函数的输入变量和输出变量的个数是可变的。

当varargin用做一个函数的输入变量时，Matlab会将其置入一个单元数组中，该数组接受由用户输入的变量数。由于varargin是一个单元数组，所以此类配置的一个重要方面是对函数的调用可包括输入的混合集。例如，若我们要使用假设函数testhv3的代码来处理此项操作，则它能很好的接受输入的混合集，如

[m, n] = testhv3(f, [0 0.5 1.5], A, 'label');

其中，f是一幅图像，下一个变量时一个长度为3的行向量，A是一个矩阵，'label'是一个字符串。

二、亮度变换的另一个M函数

在这一节中，我们将开发一个计算如下变换功能的函数：负片变换、对数变换、gamma变换和对比度拉伸变换。选用这些变换是因为随后我们将用到它们。此外，我们将进一步说明编写亮度变换M函数所涉及的机理。在编写该函数时，我们将用到函数changeclass，其语法为：

g = changeclass(newclass, f)

function image = changeclass(class, varargin) %CHANGECLASS changes the storage class of an image. %  I2 = CHANGECLASS(CLASS, I); %  RGB2 = CHANGECLASS(CLASS, RGB); %  BW2 = CHANGECLASS(CLASS, BW); %  X2 = CHANGECLASS(CLASS, X, 'indexed');   %  Copyright 1993-2002 The MathWorks, Inc.  Used with permission. %  $Revision: 1.2 $  $Date: 2003/02/19 22:09:58 $   switch class case 'uint8'    image = im2uint8(varargin{:}); case 'uint16'    image = im2uint16(varargin{:}); case 'double'    image = im2double(varargin{:}); otherwise    error('Unsupported IPT data class.'); end

此函数将图象f转换成有参数newclass指定的类别，并输出图像g。newclass的有效值是'uint8'， 'uint16'， 和'double'。

function g = intrans(f, varargin) %INTRANS Performs intensity (gray-level) transformations. %   G = INTRANS(F, 'neg') computes the negative of input image F. % %   G = INTRANS(F, 'log', C, CLASS) computes C*log(1 + F) and %   multiplies the result by (positive) constant C. If the last two %   parameters are omitted, C defaults to 1. Because the log is used %   frequently to display Fourier spectra, parameter CLASS offers the %   option to specify the class of the output as 'uint8' or %   'uint16'. If parameter CLASS is omitted, the output is of the %   same class as the input. % %   G = INTRANS(F, 'gamma', GAM) performs a gamma transformation on %   the input image using parameter GAM (a required input).  % %   G = INTRANS(F, 'stretch', M, E) computes a contrast-stretching %   transformation using the expression 1./(1 + (M./(F + %   eps)).^E).  Parameter M must be in the range [0, 1].  The default %   value for M is mean2(im2double(F)), and the default value for E %   is 4. % %   For the 'neg', 'gamma', and 'stretch' transformations, double %   input images whose maximum value is greater than 1 are scaled %   first using MAT2GRAY.  Other images are converted to double first %   using IM2DOUBLE.  For the 'log' transformation, double images are %   transformed without being scaled; other images are converted to %   double first using IM2DOUBLE. % %   The output is of the same class as the input, except if a %   different class is specified for the 'log' option.   %   Copyright 2002-2004 R. C. Gonzalez, R. E. Woods, & S. L. Eddins %   Digital Image Processing Using MATLAB, Prentice-Hall, 2004 %   $Revision: 1.7 $  $Date: 2003/10/13 00:45:53 $   % Verify the correct number of inputs. error(nargchk(2, 4, nargin))   % Store the class of the input for use later. classin = class(f);   % If the input is of class double, and it is outside the range % [0, 1], and the specified transformation is not 'log', convert the % input to the range [0, 1]. if strcmp(class(f), 'double') & max(f(:))> 1 & ...       ~strcmp(varargin{1}, 'log')    f = mat2gray(f); else % Convert to double, regardless of class(f).    f = im2double(f); end   % Determine the type of transformation specified. method = varargin{1};   % Perform the intensity transformation specified.    switch method case 'neg'    g = imcomplement(f);   case 'log'    if length(varargin) == 1        c = 1;    elseif length(varargin) == 2        c = varargin{2};    elseif length(varargin) == 3       c = varargin{2};       classin = varargin{3};    else       error('Incorrect number of inputs for the log option.')    end    g = c*(log(1 + double(f)));   case 'gamma'    if length(varargin)< 2       error('Not enough inputs for the gamma option.')    end    gam = varargin{2};    g = imadjust(f, [ ], [ ], gam);   case 'stretch'    if length(varargin) == 1       % Use defaults.       m = mean2(f);        E = 4.0;              elseif length(varargin) == 3       m = varargin{2};        E = varargin{3};    else error('Incorrect number of inputs for the stretch option.')    end    g = 1./(1 + (m./(f + eps)).^E); otherwise    error('Unknown enhancement method.') end   % Convert to the class of the input image. g = changeclass(classin, g);

例3.3 函数intrans的说明

要说明函数intrans，我们可以利用下面的例子，左边为原始图，右边为经对比度拉伸后的图像。

>> f = imread('Fig0306(a)(bone-scan-GE).tif'); >> figure(1) >> imshow(f) >> g = intrans(f, 'stretch', mean2(im2double(f)), 0.9); >> figure(2) >> imshow(g)

三、亮度标度的M函数

当处理图像是，象素值域由负到正的现象是很普遍的。尽管在中间计算过程中没有问题，但当我们想利用8bit或16bit格式保存或查看一幅图像时，就会出现问题。在这种情况下，我们通常希望吧图像标度在全尺度，即最大范围[0, 255]或[0, 65535]。下列名为gscale的M函数可以实现此功能。此外，此函数能将输出映射到一个特定的范围。

function g = gscale(f, varargin) %GSCALE Scales the intensity of the input image. %   G = GSCALE(F, 'full8') scales the intensities of F to the full %   8-bit intensity range [0, 255].  This is the default if there is %   only one input argument. % %   G = GSCALE(F, 'full16') scales the intensities of F to the full %   16-bit intensity range [0, 65535]. % %   G = GSCALE(F, 'minmax', LOW, HIGH) scales the intensities of F to %   the range [LOW, HIGH]. These values must be provided, and they %   must be in the range [0, 1], independently of the class of the %   input. GSCALE performs any necessary scaling. If the input is of %   class double, and its values are not in the range [0, 1], then %   GSCALE scales it to this range before processing. % %   The class of the output is the same as the class of the input.   %   Copyright 2002-2004 R. C. Gonzalez, R. E. Woods, & S. L. Eddins %   Digital Image Processing Using MATLAB, Prentice-Hall, 2004 %   $Revision: 1.5 $  $Date: 2003/11/21 14:36:09 $   if length(varargin) == 0 % If only one argument it must be f.    method = 'full8'; else    method = varargin{1}; end   if strcmp(class(f), 'double') & (max(f(:))> 1 | min(f(:)) < 0)    f = mat2gray(f); end   % Perform the specified scaling. switch method case 'full8'    g = im2uint8(mat2gray(double(f))); case 'full16'    g = im2uint16(mat2gray(double(f))); case 'minmax'    low = varargin{2}; high = varargin{3};    if low > 1 | low < 0 | high > 1 | high< 0       error('Parameters low and high must be in the range [0, 1].')    end    if strcmp(class(f), 'double')       low_in = min(f(:));       high_in = max(f(:));    elseif strcmp(class(f), 'uint8')       low_in = double(min(f(:)))./255;       high_in = double(max(f(:)))./255;    elseif strcmp(class(f), 'uint16')       low_in = double(min(f(:)))./65535;       high_in = double(max(f(:)))./65535;       end    % imadjust automatically matches the class of the input.    g = imadjust(f, [low_in high_in], [low high]);   otherwise    error('Unknown method.') end

调用语法为：

g = gscale(f, method, low, high)

其中，f是将被标度的图像，method的有效值为'full8'和'full16'，前者将输出标度为全范围[0, 255]，后者将输出标度为全范围[0, 65535]。若使用这两个有效值之一，则可在两种变换中省略参数low与high。method的第三个有效值为'minmax'，此时我们必须给出low与high在范围[0, 1]内的值。若选用的是'minmax'，则映射的结果