How to plot a trajectory with varying colour?

I'm trying to plot the trajectory which also shows error characteristics visually.

Inputs: East(m), North(m), error(m)

error = sqrt((East_radar - North_GPS)^2 + (East_GPS - North_GPS)^2)

At present, I'm able to plot trajectory, with the marker indices representing the points where the error is high. I have implemented the code as seen below.

To make this visualization better, how can I represent the trajectory with varying colour representing error value? as shown below.

So, higherror represents Red and low error represents Green? Your assistance would relly help me.Thanks!

If my query is unclear, please revert.

figure()
plot(Radar_East, Radar_North,'LineStyle','-')
xlabel('East (m)')
ylabel('North (m)')
title('GPS EN plot')
hold on

plot(Radar_East(outliers),Radar_North(outliers),'o','MarkerSize',8);
legend('RADAR Trajectory','Outliers > 99 percentile')
hold off

Answer 1

[Voss]

"I'm trying to assign an independent variable to a data point in 2D through colour."

You can choose any colour(s) you like.

Examples:

x = linspace(0,2*pi);
y = sin(x);

subplot(3,2,1)
c = x;
surface([x;x],[y;y],[c;c],'FaceColor','none','EdgeColor','interp','LineWidth',2);
title('color = x')

subplot(3,2,2)
c = y;
surface([x;x],[y;y],[c;c],'FaceColor','none','EdgeColor','interp','LineWidth',2);
title('color = y')

subplot(3,2,3)
c = cos(x);
surface([x;x],[y;y],[c;c],'FaceColor','none','EdgeColor','interp','LineWidth',2);
title('color = cos(x)')

subplot(3,2,4)
c = y.^2;
surface([x;x],[y;y],[c;c],'FaceColor','none','EdgeColor','interp','LineWidth',2);
title('color = y^2')

subplot(3,2,5)
c = x.^2-4*pi^2*y.^2;
surface([x;x],[y;y],[c;c],'FaceColor','none','EdgeColor','interp','LineWidth',2);
title('color = x^2-4*pi^2*y^2')

subplot(3,2,6)
c = rand(size(x));
surface([x;x],[y;y],[c;c],'FaceColor','none','EdgeColor','interp','LineWidth',2);
title('color = random')

figure
colormap(copper())

subplot(3,2,1)
c = x;
surface([x;x],[y;y],[c;c],'FaceColor','none','EdgeColor','interp','LineWidth',2);
title('color = x')

subplot(3,2,2)
c = y;
surface([x;x],[y;y],[c;c],'FaceColor','none','EdgeColor','interp','LineWidth',2);
title('color = y')

subplot(3,2,3)
c = cos(x);
surface([x;x],[y;y],[c;c],'FaceColor','none','EdgeColor','interp','LineWidth',2);
title('color = cos(x)')

subplot(3,2,4)
c = y.^2;
surface([x;x],[y;y],[c;c],'FaceColor','none','EdgeColor','interp','LineWidth',2);
title('color = y^2')

subplot(3,2,5)
c = x.^2-4*pi^2*y.^2;
surface([x;x],[y;y],[c;c],'FaceColor','none','EdgeColor','interp','LineWidth',2);
title('color = x^2-4*pi^2*y^2')

subplot(3,2,6)
c = rand(size(x));
surface([x;x],[y;y],[c;c],'FaceColor','none','EdgeColor','interp','LineWidth',2);
title('color = random')

[Maheedhar Korimi]

Thank you @Voss for the explanation, but in this case y is a function of x.

Here I have the coordinates of the object (x,y) and would like to measure error in comparison with ground truth. So this method unfortunately does not work. can you please refer to the files attached in the previous comment?

Thanks for the effort.

[Voss]

Your variables are column vectors; the variables in my examples were row vectors. Easy enough to make it work. See below.

load('test_files_Data.mat')

Error = sqrt((RPS_E-GPS_E).^2+(RPS_N-GPS_N).^2);
mean_Error = mean(Error,1);
Error_std = Error - mean_Error;

x = GPS_E;
y = GPS_N;
c = Error_std;
surface([x x],[y y],[c c],'FaceColor','none','EdgeColor','interp','LineWidth',2);

colormap(jet());
cb = colorbar();
ylabel(cb,'"Error_std"','Interpreter','none')

[Maheedhar Korimi]

That's what I'm looking for, thank you for the help, that was a silly mistake on my part.

Cheers!!

Thank you @Voss

[Voss]

You're welcome!

Answer 2

[Cameron]

I don't think there is a way to do this using just one plot. My quick solution would be to do something like this

x = 0:0.05:2*pi; %sample x data
y = sin(x); %sample y data
c = hsv(length(x)); %hsv values over the length of x
hold on %hold the graph
for ii = 1:length(x)-1
    plot(x(ii:ii+1),y(ii:ii+1),'-','Color',c(ii,:)) %plot
end
hold off 

Not the most elegant solution, but if really want to do it, this is a way. One problem with this approach would be if you have few data points. You could spline or smooth your data though.

[John D'Errico]

My gut said oh, sure. This must be possible, probably using line. Line should accept a color spec at each point along the line. It just makes sense this is where I would find it, with one of the possibiities for the 'color' property would be the colors to use. Then I checked. Time for a feature request. :)

[Maheedhar Korimi]

@Cameron Thanks for the quick response. Due to confidentiality, I couldn't share the actual trajectory (which covers an area of 9 square Kilometer). So the plot is huge and contains many data points.

In your solution, I cannot assign colour to the error value, so I cannot visualize the error properly, so I don't think this is the right solution.

Appreciate your effort, Thanks

[Maheedhar Korimi]

@John D'Errico Thanks for understanding my question i.e., using colors to represent underlying data is what I'm looking for.

Cheers for the comment.

Answer 3

[Les Beckham]

Take a look at this Answer which shows a trick for doing this using surface.

https://www.mathworks.com/matlabcentral/answers/5042-how-do-i-vary-color-along-a-2d-line#answer_7057

[Maheedhar Korimi]

I have tried this initially, but unfortunately couldn't work with it. I'm trying to assign an independent variable to a data point in 2D through colour.

[Les Beckham]

If you provide more details (your data and the code you are using to plot it, and the exact errors or undesired behavior you are seeing) someone can probably help work out the issues you are having.

[Maheedhar Korimi]

Please refer the attachments for data! My apologies for the delay.

Thank you

[Les Beckham]

"provide more details (your data and the code you are using to plot it".

Are we supposed to guess what you are trying to plot (and how)?

[Maheedhar Korimi]

I have attached the code and data in my previous comment, please run it, you'll see a trajectory(radar) with errors (> 99 percentile) in comparision with GPS data, represented as marker indices.

So, instead of representing the errors as marker indices, I would like to show the errors in the form of colour (maybe of colourmap: jet), where blue represents low error and red represents high error in the data.

If you have any further questions, please feel free to ask.

My apologies if I haven't communicated properly.