matlab中imagesc函数得到图像，怎么还原？

Posted 2023-05-12

参考技术A 理论上从图像是不能完全恢复出原来的矩阵数据的
因为在由数据转化为图像的过程中量化的时候丢失了信息
可以从得到的图像和所使用的colormap恢复一定的数据信息
如果得到图像的数据假设为变量im，应该是1个mxnx3的矩阵
m和n表示图像的像素，第三维大小为3，分别储存图像的r，g，b数据
又知道原来的画图所用的colormap，假设为变量map
那么可以用函数a=
rgb2ind(im,map)
这时候得到的a是一个整数矩阵
b=double(a)./double(max(a(:)));
这样得到的b是一个数值范围从0到1的矩阵
0表示最小值，1表示最大值
这是从图像恢复数据可以做到的程度，和原来的数据有较大差距
如果你知道原来的数据的数值范围，也就是最小值和最大值minv
maxv
那么还可以
c=minv+(maxv-minv)*b;
进一步得到和原数据相似的矩阵，但是已经不能100%还原矩阵了

MATLAB 颜色图函数（imagesc/scatter/polarPcolor/pcolor）

2维的热度图 imagesc

imagesc（x, y, z），x和y分别是横纵坐标，z为值，表示颜色

1 imagesc(theta,phi,slc); colorbar
2 xlabel(‘theta(°)‘,‘fontname‘,‘Times New Roman‘,‘FontSize‘,14);
3 ylabel(‘phi(°)‘,‘fontname‘,‘Times New Roman‘,‘FontSize‘,14);
4 sta = ‘3 objects at (θ,φ,r) : (-30,30,1) (0,0,2) (60,-60,0.5)‘;
5 str=sprintf(strcat(‘3D Imaging Slice at :‘, num2str(d_max*D/N), ‘(m)‘, ‘
‘,sta));
6 title(str, ‘fontname‘,‘Times New Roman‘,‘Color‘,‘k‘,‘FontSize‘,13);
7 grid on

 

 

其中，colorbar的坐标值调整：caxis([0 1]);

 colormap的色系调整：colormap hot

 

3维散点图 scatter

1 scatter3(x,y,z,24,c,‘filled‘);
2 % axis([-(R+2) (R+2) -(R+2) (R+2) 0 (h+2)]);
3 colorbar

 

 

2维 极坐标热度图 polarPcolor

polarPcolor(R_axis, theta, value)，前两个为半径方向坐标轴和圆心角坐标轴，value为值，用颜色表示

1 [fig, clr] = polarPcolor(R_axis, theta, x_d_th, ‘labelR‘,‘range (m)‘,‘Ncircles‘, 5,‘Nspokes‘,7);
2 colormap hot
3 % caxis([0 1]);

 

 

其中polarPcolor代码如下：

  1 function [varargout] = polarPcolor(R,theta,Z,varargin)
  2 % [h,c] = polarPcolor1(R,theta,Z,varargin) is a pseudocolor plot of matrix 
  3 % Z for a vector radius R and a vector angle theta. 
  4 % The elements of Z specify the color in each cell of the 
  5 % plot. The goal is to apply pcolor function with a polar grid, which 
  6 % provides a better visualization than a cartesian grid.
  7 %
  8 %% Syntax
  9 % 
 10 % [h,c] = polarPcolor(R,theta,Z)
 11 % [h,c] = polarPcolor(R,theta,Z,‘Ncircles‘,10)
 12 % [h,c] = polarPcolor(R,theta,Z,‘Nspokes‘,5)
 13 % [h,c] = polarPcolor(R,theta,Z,‘Nspokes‘,5,‘colBar‘,0) 
 14 % [h,c] = polarPcolor(R,theta,Z,‘Nspokes‘,5,‘labelR‘,‘r (km)‘)
 15 % 
 16 % INPUT
 17 %    * R :
 18 %        - type: float
 19 %        - size: [1 x Nrr ] where Nrr = numel(R).
 20 %        - dimension: radial distance.
 21 %    * theta : 
 22 %        - type: float
 23 %        - size: [1 x Ntheta ] where Ntheta = numel(theta).
 24 %        - dimension: azimuth or elevation angle (deg).
 25 %        - N.B.: The zero is defined with respect to the North.
 26 %    * Z : 
 27 %        - type: float
 28 %        - size: [Ntheta x Nrr]
 29 %        - dimension: user‘s defined .
 30 %    * varargin:
 31 %        - Ncircles: number  of circles for the grid definition.
 32 %        - Nspokes: number of spokes for the grid definition.
 33 %        - colBar: display the colorbar or not.
 34 %        - labelR: legend for R.
 35 % 
 36 % 
 37 % OUTPUT
 38 % h: returns a handle to a SURFACE object.
 39 % c: returns a handle to a COLORBAR object.
 40 %
 41 %% Examples 
 42 % R = linspace(3,10,100);
 43 % theta = linspace(0,180,360);
 44 % Z = linspace(0,10,360)‘*linspace(0,10,100);
 45 % figure
 46 % polarPcolor(R,theta,Z,‘Ncircles‘,3)
 47 %
 48 %% Author
 49 % Etienne Cheynet, University of Stavanger, Norway. 28/05/2016
 50 % see also pcolor
 51 % 
 52 
 53 %%  InputParseer
 54 p = inputParser();
 55 p.CaseSensitive = false;
 56 p.addOptional(‘Ncircles‘,5);
 57 p.addOptional(‘Nspokes‘,8);
 58 p.addOptional(‘labelR‘,‘‘);
 59 p.addOptional(‘colBar‘,1);
 60 p.parse(varargin{:});
 61 
 62 Ncircles = p.Results.Ncircles ;
 63 Nspokes = p.Results.Nspokes ;
 64 labelR = p.Results.labelR ;
 65 colBar = p.Results.colBar ;
 66 %% Preliminary checks
 67 % case where dimension is reversed
 68 Nrr = numel(R);
 69 Noo = numel(theta);
 70 if isequal(size(Z),[Noo,Nrr]),
 71     Z=Z‘;
 72 end
 73 
 74 % case where dimension of Z is not compatible with theta and R
 75 if ~isequal(size(Z),[Nrr,Noo])
 76     fprintf(‘
‘)
 77     fprintf([ ‘Size of Z is : [‘,num2str(size(Z)),‘] 
‘]);
 78     fprintf([ ‘Size of R is : [‘,num2str(size(R)),‘] 
‘]);
 79     fprintf([ ‘Size of theta is : [‘,num2str(size(theta)),‘] 

‘]);
 80     error(‘ dimension of Z does not agree with dimension of R and Theta‘)
 81 end
 82 %% data plot
 83 rMin = min(R);
 84 rMax = max(R);
 85 thetaMin=min(theta);
 86 thetaMax =max(theta);
 87 % Definition of the mesh
 88 Rrange = rMax - rMin; % get the range for the radius
 89 rNorm = R/Rrange; %normalized radius [0,1]
 90 % get hold state
 91 cax = newplot;
 92 % transform data in polar coordinates to Cartesian coordinates.
 93 YY = (rNorm)‘*cosd(theta);
 94 XX = (rNorm)‘*sind(theta);
 95 % plot data on top of grid
 96 h = pcolor(XX,YY,Z,‘parent‘,cax);
 97 shading flat
 98 set(cax,‘dataaspectratio‘,[1 1 1]);axis off;
 99 if ~ishold(cax);
100     % make a radial grid
101     hold(cax,‘on‘)
102     % Draw circles and spokes
103     createSpokes(thetaMin,thetaMax,Ncircles,Nspokes);
104     createCircles(rMin,rMax,thetaMin,thetaMax,Ncircles,Nspokes)
105 end
106 
107 %% PLot colorbar if specified
108 if colBar==1,
109     c =colorbar(‘location‘,‘WestOutside‘);
110     caxis([quantile(Z(:),0.01),quantile(Z(:),0.99)])
111 else
112     c = [];
113 end
114 
115 %% Outputs
116 nargoutchk(0,2)
117 if nargout==1,
118     varargout{1}=h;
119 elseif nargout==2,
120     varargout{1}=h;
121     varargout{2}=c;
122 end
123 
124 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
125 % Nested functions
126 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
127     function createSpokes(thetaMin,thetaMax,Ncircles,Nspokes)
128         
129         circleMesh = linspace(rMin,rMax,Ncircles);
130         spokeMesh = linspace(thetaMin,thetaMax,Nspokes);
131         contour = abs((circleMesh - circleMesh(1))/Rrange+R(1)/Rrange);
132         cost = cosd(90-spokeMesh); % the zero angle is aligned with North
133         sint = sind(90-spokeMesh); % the zero angle is aligned with North
134         for kk = 1:Nspokes
135             plot(cost(kk)*contour,sint(kk)*contour,‘k:‘,...
136                 ‘handlevisibility‘,‘off‘);
137             % plot graduations of angles
138             % avoid superimposition of 0 and 360
139             if and(thetaMin==0,thetaMax == 360),
140                 if spokeMesh(kk)<360,
141                     
142                     text(1.05.*contour(end).*cost(kk),...
143                         1.05.*contour(end).*sint(kk),...
144                         [num2str(spokeMesh(kk),3),char(176)],...
145                         ‘horiz‘, ‘center‘, ‘vert‘, ‘middle‘);
146                 end
147             else
148                 text(1.05.*contour(end).*cost(kk),...
149                     1.05.*contour(end).*sint(kk),...
150                     [num2str(spokeMesh(kk),3),char(176)],...
151                     ‘horiz‘, ‘center‘, ‘vert‘, ‘middle‘);
152             end
153             
154         end
155     end
156     function createCircles(rMin,rMax,thetaMin,thetaMax,Ncircles,Nspokes)
157         
158         % define the grid in polar coordinates
159         angleGrid = linspace(90-thetaMin,90-thetaMax,100);
160         xGrid = cosd(angleGrid);
161         yGrid = sind(angleGrid);
162         circleMesh = linspace(rMin,rMax,Ncircles);
163         spokeMesh = linspace(thetaMin,thetaMax,Nspokes);
164         contour = abs((circleMesh - circleMesh(1))/Rrange+R(1)/Rrange);
165         % plot circles
166         for kk=1:length(contour)
167             plot(xGrid*contour(kk), yGrid*contour(kk),‘k:‘);
168         end
169         % radius tick label
170         for kk=1:Ncircles
171             
172             position = 0.51.*(spokeMesh(min(Nspokes,round(Ncircles/2)))+...
173                 spokeMesh(min(Nspokes,1+round(Ncircles/2))));
174             
175             if abs(round(position)) ==90,
176                 % radial graduations
177                 text((contour(kk)).*cosd(90-position),...
178                     (0.1+contour(kk)).*sind(86-position),...
179                     num2str(circleMesh(kk),2),‘verticalalignment‘,‘BaseLine‘,...
180                     ‘horizontalAlignment‘, ‘center‘,...
181                     ‘handlevisibility‘,‘off‘,‘parent‘,cax);
182                 
183                 % annotate spokes
184                 text(contour(end).*0.6.*cosd(90-position),...
185                     0.07+contour(end).*0.6.*sind(90-position),...
186                     [labelR],‘verticalalignment‘,‘bottom‘,...
187                     ‘horizontalAlignment‘, ‘right‘,...
188                     ‘handlevisibility‘,‘off‘,‘parent‘,cax);
189             else
190                 % radial graduations
191                 text((contour(kk)).*cosd(90-position),...
192                     (contour(kk)).*sind(90-position),...
193                     num2str(circleMesh(kk),2),‘verticalalignment‘,‘BaseLine‘,...
194                     ‘horizontalAlignment‘, ‘right‘,...
195                     ‘handlevisibility‘,‘off‘,‘parent‘,cax);
196                 
197                 % annotate spokes
198                 text(contour(end).*0.6.*cosd(90-position),...
199                     contour(end).*0.6.*sind(90-position),...
200                     [labelR],‘verticalalignment‘,‘bottom‘,...
201                     ‘horizontalAlignment‘, ‘right‘,...
202                     ‘handlevisibility‘,‘off‘,‘parent‘,cax);
203             end
204         end
205         
206     end
207 end

View Code

再贴一个示例代码：

 1 %% Examples
 2 % The following examples illustrate the application of the function 
 3 % polarPcolor
 4 clearvars;close all;clc;
 5 
 6 %% Minimalist example
 7 % Assuming that a remote sensor is measuring the wind field for a radial
 8 % distance ranging from 50 to 1000 m. The scanning azimuth is oriented from
 9 % North (0 deg) to North-North-East ( 80 deg):
10 R = linspace(50,1000,100)./1000; % (distance in km)
11 Az = linspace(0,80,100); % in degrees
12 [~,~,windSpeed] = peaks(100); % radial wind speed
13 figure(1)
14 [h,c]=polarPcolor(R,Az,windSpeed);
15 
16 %% Example with options
17 %  We want to have 4 circles and 7 spokes, and to give a label to the
18 %  radial coordinate
19 
20 figure(2)
21 [~,c]=polarPcolor(R,Az,windSpeed,‘labelR‘,‘r (km)‘,‘Ncircles‘,7,‘Nspokes‘,7);
22 ylabel(c,‘ radial wind speed (m/s)‘);
23 set(gcf,‘color‘,‘w‘)
24 %% Dealing with outliers
25 % We introduce outliers in the wind velocity data. These outliers
26 % are represented as wind speed sample with a value of 100 m/s. These
27 % corresponds to unrealistic data that need to be ignored. To avoid bad
28 % scaling of the colorbar, the function polarPcolor uses the function caxis
29 % combined to the function quantile to keep the colorbar properly scaled:
30 % caxis([quantile(Z(:),0.01),quantile(Z(:),0.99)])
31 
32 windSpeed(1:10:end,1:20:end)=100;
33 
34 figure(3)
35 [~,c]=polarPcolor(R,Az,windSpeed);
36 ylabel(c,‘ radial wind speed (m/s)‘);
37 set(gcf,‘color‘,‘w‘)
38 
39 %% polarPcolor without colorbar
40 % The colorbar is activated by default. It is possible to remove it by 
41 % using the option ‘colBar‘. When the colorbar is desactivated, the
42 % outliers are not "removed" and bad scaling is clearly visible:
43 
44 figure(4)
45 polarPcolor(R,Az,windSpeed,‘colBar‘,0) ;
46 
47 
48 %% Different geometry 1
49 N = 360;
50 R = linspace(0,1000,N)./1000; % (distance in km)
51 Az = linspace(0,360,N); % in degrees
52 [~,~,windSpeed] = peaks(N); % radial wind speed
53 figure(5)
54 [~,c]= polarPcolor(R,Az,windSpeed);
55 ylabel(c,‘ radial wind speed (m/s)‘);
56 set(gcf,‘color‘,‘w‘)
57 %% Different geometry 2
58 N = 360;
59 R = linspace(500,1000,N)./1000; % (distance in km)
60 Az = linspace(0,270,N); % in degrees
61 [~,~,windSpeed] = peaks(N); % radial wind speed
62 figure(6)
63 [~,c]= polarPcolor(R,Az,windSpeed,‘Ncircles‘,3);
64 location = ‘NorthOutside‘;
65 ylabel(c,‘ radial wind speed (m/s)‘);
66 set(c,‘location‘,location);
67 set(gcf,‘color‘,‘w‘)

View Code