How to define the boundary of a countour

I am using this line of code to plot my contour of sea surface temperature, it is a rectangular box. [C1, h1] = contourf (X, Y, Z, [0:1:35], 'LineStyle', 'none' );

How do I trim off the areas that extends to the land with a polygon?

I tried to use "patch", but it does not work well, when the plot from the Atlantic Ocean is put together with the plots from other ocean basins.

Is there a way I can set X, Y, Z values (gridded data) to be NaN for the grids that are outside the polygon?

Thank you.

[Image Analyst]

Does this require the Mapping Toolbox? If so, list it below in the "Products" section.

[Leon]

Thank you for the replies!

No. I plotted the map using data that I downloaded somewhere else.

[Image Analyst]

Is your data an RGB image? Or is it in some other form? And explain "areas that extends to the land" - how do we know which pixels are land or not land? Do you have a binary image that defines land/not-land pixels?

[Leon]

You mean the map data? No, they are just two columns of longitude and latitude.

I have a polygon of the boundary of Atlantic Ocean. So the function "inpolygon" only works for column data, not grid data?

Answer 1

[Chad Greene]

If you have the mapping toolbox you can use landmask to set land values to NaN.

[Leon]

This is great! Thank you so much.

Is there a way I can trim more values off? For example, the areas that extends to the Pacific Ocean.

[Chad Greene]

You can simply set the limits of the map when you call worldmap. Below I set the limits from 75S to 75N and 80W to 20E.

[lon,lat] = meshgrid(linspace(-179, 179, 180),...
    linspace(-89.5,89.5, 180));
z = peaks(180);
z(landmask(lat,lon))=NaN; 

worldmap([-75 75],[-80 20])
pcolorm(lat,lon,z)

Answer 2

[Image Analyst]

Do you have the Image Processing Toolbox? If so, you can use poly2mask() to create a mask of land/not-land. Then you can mask the image.

% Mask the image using bsxfun() function
maskedRgbImage = bsxfun(@times, rgbImage, cast(mask, class(rgbImage)));

[Leon]

Thank you. Unfortunately, I do not have image processing toolbox.

[Image Analyst]

Then you can create a binary mask with a double for loop. You can still use bsxfun() to do the masking though - that does not require the toolbox.