信噪比图和矩形拟合（Matlab） 我附上了3幅图像和信噪比函数来计算每幅图像之间的信噪比。我如何绘制这个信噪比，以便从图中容易理解电平，而不仅仅是数字

信噪比函数：

RMS函数

剧本

% sig= best focussed image
% noisy= off focussed image
% dark = no light image
%-------------------------
% calculate SNR between:
% sig and noise
% signal and dark
% noise and dark
clear
sig = rgb2gray(imread('S1-BestFocus.bmp'));
noisy = rgb2gray(imread('S1-OffFocus.bmp'));
dark=rgb2gray(imread('DarkScan.bmp'));
sig_noise = SNR(sig,noisy,'db',true);
sig_dark = SNR(sig,dark,'db',true);
noise_dark = SNR(noisy,dark,'db',true);

数字：

我正在成像一个宽度为15-18微米、长度为1mm的狭缝，狭缝不均匀，因此我必须检查狭缝长度的宽度变化有多大。获得测量值的最佳方法是什么。（一种方法是使用矩形拟合）

function RMS= rms(varargin)
%
% Written by Phillip M. Feldman  March 31, 2006
%
% rms computes the root-mean-square (RMS) of values supplied as a
% vector, matrix, or list of discrete values (scalars).  If the input is
% a matrix, rms returns a row vector containing the RMS of each column.

% David Feldman proposed the following simpler function definition:
%
%    RMS = sqrt(mean([varargin{:}].^2))
%
% With this definition, the function accepts ([1,2],[3,4]) as input,
% producing 2.7386 (this is the same result that one would get with
% input of (1,2,3,4).  I'm not sure how the function should behave for
% input of ([1,2],[3,4]).  Probably it should produce the vector
% [rms(1,3) rms(2,4)].  For the moment, however, my code simply produces
% an error message when the input is a list that contains one or more
% non-scalars.

if (nargin == 0)
   error('Missing input.');
end


% Section 1: Restructure input to create x vector.

if (nargin == 1)
   x= varargin{1};

else

   for i= 1 : size(varargin,2)
      if (prod(size(varargin{i})) ~= 1)
         error(['When input is provided as a list, ' ...
                'list elements must be scalar.']);
      end

      x(i)= varargin{i};
   end
end


% Section 2: Compute RMS value of x.

RMS= sqrt (mean (x .^2) );

% sig= best focussed image
% noisy= off focussed image
% dark = no light image
%-------------------------
% calculate SNR between:
% sig and noise
% signal and dark
% noise and dark
clear
sig = rgb2gray(imread('S1-BestFocus.bmp'));
noisy = rgb2gray(imread('S1-OffFocus.bmp'));
dark=rgb2gray(imread('DarkScan.bmp'));
sig_noise = SNR(sig,noisy,'db',true);
sig_dark = SNR(sig,dark,'db',true);
noise_dark = SNR(noisy,dark,'db',true);