ASF - Alternate Sequential Filtering.

Description
Syntax
Inputs
Outputs
References
See also
Function implementation

Description

Perform the Alternate Sequential Filtering (ASF) and the ASF by reconstruction of an image [Serra88,Soille03].

Syntax

   F = ASF(I);
   F = ASF(I, op, n);
   F = ASF(I, op, n, shape);
   F = ASF(I, op, n, shape, s1);
   F = ASF(I, op, n, shape, s1, s2);

Inputs

I : input image, possibly multispectral.

op : string defining if the filter is by reconstruction (it should start with a 'r' then), and specifying the order in which the opening ('o') and closing ('c') operators are applied; it can be either:

'oc' or 'co' for standard ASF applying resp. first the opening, then the closing, or first the opening, then the closing,
'oco' or 'coc' for standard ASF applying resp., in this order, opening, closing and opening or closing, opening and closing,
'rco' or 'roc' for ASF by reconstruction applying resp. first the opening, then the closing, or first the closing, then the opening,
'roco' or 'rcoc' for ASF by reconstruction applying resp., in this order, opening, closing and opening or closing, opening and closing,

default: op='oc'.

n : number of iterations of the sequential filter; default: n=1.

shape : any string among 'disk', 'rectangle', 'square', 'diamond', 'line', 'periodicline' and 'octagon' (in particular, 'ball', 'arbitrary' and 'pair' are not supported) or any scalar value. In the former case, a default SE of type shape and size 3 will be output, in the latter case, a SE of type 'disk' and size shape will be output; default: shape='disk'.

s1 : optional shape (string, see above) or size of the SE when nargin=2; when nargin=3 (ie. s2 is also defined), s1 is necessarly a scalar value.

s2 : optional scalar value for defining the SE (see STREL possibly defined with 2 parameters: 'periodicline', 'disk', 'line').

Outputs

F : output of the ASF or the ASF by reconstruction applied on the input image.

References

[Serra88] J. Serra: "Image Analysis and Mathematical Morphology", vol. 2, Academic Press, New York (USA), 1988.

[Soille03] P. Soille: "Morphological Image Analysis?Principles and Applications", 2nd ed. Berlin (Germany), Springer-Verlag, 2003.

Function implementation

function F = asf(I, varargin)

if isempty(ver('images'))
    error('asf:errortoolbox', 'Image Processing toolbox required');
end

error(nargchk(1, 17, nargin, 'struct'));
error(nargoutchk(1, 1, nargout, 'struct'));

parsing parameters

if ~isnumeric(I)
    error('asf:inputparameter','a matrix is required in input');
end

% optional parameters
p = createParser('ASF');

p.addOptional('op', 'oc', @(x)ischar(x) && ...
    any(strcmpi(x,{'co','oc','oco','coc','rco','roc','rcoc','roco'})));
p.addOptional('n', 1, @(x)isnumeric(x) && x>=1);
p.addOptional('shape', 'disk', @(x) ischar(x) && ...
    any(strcmpi(x,{'disk','rectangle','square','diamond', ...
                'line','periodicline','octagon'})));
p.addParamValue('s1', 3, @(x)isnumeric(x));
p.addParamValue('s2', [], @(x)isnumeric(x));

% parse and validate all input arguments
p.parse(varargin{:});
p = getvarParser(p);

main processing

F = asf_base(I, p.n, p.shape, p.op, p.s1, p.s2);

display

if p.disp
    thetitle = 'ASF';
    if p.op(1)=='r', thetitle = [thetitle ' by reconstruction'];  end
    thetitle = [thetitle ' - order: ' p.op(end-1:end)];
    figure, imagesc(rescale(F)), axis image off, title(thetitle);
    if size(I,3)==1,  colormap gray;  end
end

end % end of asf