Superimposing images (maps) with different colour scales and transparency/masking?

I am trying to plot two maps (same size) on top of each other, where the first one is in colour scale and the second one is in gray-scale, semitransparent (or hatched pattern)

I would like to display input_A as a regular map with colour scale jet(15). Next, I would like to superimpose input_B of the same size and on the same axis, which consists of only two values (0 and 1), plotted in semi-transparent gray for cell values of 1 only.

(If the value of input_B is 0, the cell should be fully transparent (showing only input_A in the end plot) and if the cell value of input_B is 1, it should be semi-transparent fill, so that on the final figure shows input_A with a dot shaded in gray from input_B).

   %To plot map_A, I’m using the code:

colour_map_A=image(min(longitude):3.6:max(longitude),max(latitude):-1.8:min(latitude),input_A,'CDataMapping','scaled')
  caxis(scale_A) %scale_A=[-100,100];
  colormap(flipud(jet(15)))
  axis equal
  box on
  set(gca,'XLim',[min(longitude) max(longitude],'YLim',[min(latitude) max(latitude)], ...
        'XTick',[0 60 120 180 240 300 360], ...
        'Ytick',[-90 -60 -30 0 30 60 90]);
  xlabel('Longitude')
  ylabel('Latitude')
  % load topographic data
  load('topo.mat','topo');
  % plot coast lines
  hold on
  contour(0:359,90:-1:-89,topo,[0,0],'k')

hold on

gray_map_B=image(min(longitude):3.6:max(longitude),max(latitude):-1.8:min(latitude),input_B,'CDataMapping','scaled')
  caxis(scale_B) %scale_B=[0,1];
  colormap(flipud(gray(2)))

alpha=0.3; %however, this makes both plots semi-transparent, not only the second one.

Also, if I plot gray_map_B on top of colour_map_A, the colour scale changes for both maps to gray and gray_map_B covers colour_map_A.

In the end, I’m looking for a way of plotting input_A as it is in colour map, and superimposing input_B in a semi-transparent gray only for cell values of 1 only.

[Image Analyst]

Attach input_A, input_B, and the 'topo.mat' file so people can try to help you.

Answer 1

[Walter Roberson]

Bimage = get(gray_map_B, 'CData');
set(gray_map_B, 'AlphaData', 0.3 * (Bimage == 1));

[k78na]

thank you, it seems to work for part of my question, but it uses the colormap of the original image and fades it out for some cells, as needed. However, I would like to create a faded "mask" of a different colour (semi-transparent, on top of the original image), rather than shading the original image.

I tried

set(gray_map_B, 'AlphaData', 0.3 * (Bimage == 1));
hold on;
set(gray_map_B, 'Colormap',mycmap_gray)

(or I could just use build-in colour map gray(2) to have gray shading on the mask)

However, this gives an error:

Error using matlab.graphics.primitive.Image/set
There is no Colormap property on the Image class.

On the other hand, using

    set(figure(1),'Colormap',mycmap_gray)

changes color maps of both images, while I only wanted to change the colour of the created mask (gray_map_B).

Would you have any suggestions on how to set different colormap for gray_map_B, without impacting the colour_map_A?

Answer 2

[Image Analyst]

Does either of these explanations help?

http://blogs.mathworks.com/steve/2009/02/18/image-overlay-using-transparency/

http://www.mathworks.com/matlabcentral/answers/104889-how-to-overlay-a-foreground-image-over-a-background-image

[k78na]

Thank you, the bottom of the page of the first link is helpful. However, how to make the green mask semi-transparent (and the transparency depends on the cell value of I)?

E=input_A; %main image
I=input_B; %mask

green = cat(3, zeros(size(E)), ones(size(E)), zeros(size(E)));
hold on

h=image(min(longitude):3.6:max(longitude),max(latitude):-1.8:min(latitude),green,'CDataMapping','scaled')
hold off

%set(h, 'AlphaData', I) 

%this works, but the mask is fully non-transparent and the values of input_A below the mask are not visible

I tried following advice from the previous answer to make the green mask semi-transparent, but the information from mask I is not included then

Bimage = get(h, 'CData');
set(h, 'AlphaData', 0.1 *(Bimage == 1));

and it gives the error:

While setting the 'AlphaData' property of Image:
Value must be a scalar, vector or array of numeric or logical type

Please note that Bimage is a matrix with values of 0 and 1 only, and so is I. The values of 0 should be fully transparent, while the values of 1 should be semi-transparent, so that the original image is still visible below the mask).

setting alpha=0.3; does not help, as it also changes the transparency of the original image E, which should not be affected by alpha.

(Also, I need to use image(...) instead of imshow(...) as I'm plotting maps with topographic data and coast lines).

Would you have any suggestions how to make the mask semi-transparent, according to values of I?

[Harry Farmer]

Hi,

I was wondering if you ever managed to resolve the "'AlphaData' property of Image: Value must be a scalar, vector or array of numeric or logical type" error you were getting. I'm trying to do something quite a bit more simple than you (just making one image transparent using another image's alpha mask, but am getting this message and can't find anything on the internet that would indicate why. Infact yours is the only report of this error I can find online.

[Walter Roberson]

In the above code, Bimage is 3D, so Bimage == 1 is 3D, and AlphaData cannot be 3D. The above code should use

set(h, 'AlphaData', 0.1 * (size(Bimage,1), size(Bimage,2)));