Colorbar with background image

Hello!

I am trying to plot some short line segments, colored by time. This works well using surf. However, I want to also have a background binary image (here, "BW") that I am plotting these segments over.

I am running into an issue where the background image changes the scaling of the colorbar, which I want to be scaled to time. I think in theory I could use clim, but I am not sure how to isolate the axes of just my segments (and not BW) since I am using surf in a for loop.

Any help would be greatly appreciated!

Thanks,

VK

% example code
figure; hold on
BW = (imbinarize(image));
imshow(BW); % background image
for i = 1:length(tracks) % plot each segment
    x = [tracks(i).x]; y = [tracks(i).y]; z = [tracks(i).time];
    s = surf([x(:) x(:)],[y(:) y(:)], [z(:) z(:)],...
    'FaceColor','none',...
    'EdgeColor','interp',...
    'LineWidth',2);
end 
view(2)
colorbar % very close, just need to apply the colors to the segments only and not background image...

[Image Analyst]

How about a screenshot of what you got and what you want?

[dpb]

I would overlay two axes and then you can put the background image in the one and the surf() components in the other -- and then pass the object handles to colorbar

[Victoria]

Yep, totally spaced on including some screenshots!

The first image is what I am currently getting, while the second is what I want (minus the coloring by time in each segment; these segements are just randomly colored and don't use surf, but instead plot with hold on).

I guess I am stuck on how to save the axes from an "updating" surf (since it's in a for loop with hold on). Trying the code below only spits out the very last segment in ax2. Maybe there is a different way to color each segment by time instead of using surf?

figure; hold on
BW = squeeze(imbinarize(arr_0(1,:,:)));
imshow(BW); % background image
ax1=gca;
ax2 = axes(Position=ax1.Position);
for i = 1:length(tracksToUse) % for each segment
    x = [tracksToUse(i).xLoc]/99; y = [tracksToUse(i).yLoc]/99; z = [tracksToUse(i).frame];
    surf(ax2,[x(:) x(:)],[y(:) y(:)], [z(:) z(:)],...
    'FaceColor','none',...
    'EdgeColor','interp',...
    'LineWidth',2);
end 

[Matt J]

Maybe there is a different way to color each segment by time instead of using surf?

Use surface instead of surf. And don't use hold on.

figure; %hold on
BW = squeeze(imbinarize(arr_0(1,:,:)));
imshow(BW); % background image
ax1=gca;
ax2 = axes(Position=ax1.Position);
for i = 1:length(tracksToUse) % for each segment
    x = [tracksToUse(i).xLoc]/99; y = [tracksToUse(i).yLoc]/99; z = [tracksToUse(i).frame];
    surface(ax2,[x(:) x(:)],[y(:) y(:)], [z(:) z(:)],...
    'FaceColor','none',...
    'EdgeColor','interp',...
    'LineWidth',2);
end 

Or you can use hold on, but apply it to ax2,

figure;
BW = squeeze(imbinarize(arr_0(1,:,:)));
imshow(BW); % background image
ax1=gca;
ax2 = axes(Position=ax1.Position);

hold(ax2,'on');
for i = 1:length(tracksToUse) % for each segment
    x = [tracksToUse(i).xLoc]/99; y = [tracksToUse(i).yLoc]/99; z = [tracksToUse(i).frame];
    surf(ax2,[x(:) x(:)],[y(:) y(:)], [z(:) z(:)],...
    'FaceColor','none',...
    'EdgeColor','interp',...
    'LineWidth',2);
end 

All of this is incidental, however, to your original question about colorbars. To add colorbars, see my answer below.

Answer 1

[Matt J]

You can use this as a template:

% Create sample data
[X,Y] = meshgrid(linspace(-3,3,200));
A = exp(-(X.^2 + Y.^2));          % smooth Gaussian
B = sin(3*X).*cos(3*Y);           % oscillatory pattern


% Plot both in overlapping axes

h1 = imagesc(A);
ax1=gca; 
ax2 = axes(Position=ax1.Position); 
h2 = surf(ax2,X,Y, B,'FaceAlpha',0.05,'EdgeColor','none');      

colormap(ax1, gray(256))
colormap(ax2, 'parula')

axis([ax1,ax2],'off');
colorbar(ax1,'Visible','off');    
colorbar(ax2);           

title({'Surface B overlayed on image A', 'Colorbar corresponds to B'});
set(gcf,'Position',gcf().Position.*[1,1,1,1.2])

Answer 2

[Voss]

tmp = arrayfun(@(x)x.time(:),tracks,'UniformOutput',false);
tmp = vertcat(tmp{:});
[z_min,z_max] = bounds(tmp);
clim([z_min,z_max])