{"id":260,"date":"2009-05-15T07:00:57","date_gmt":"2009-05-15T11:00:57","guid":{"rendered":"https:\/\/blogs.mathworks.com\/steve\/2009\/05\/15\/continental-divide-4-oceans\/"},"modified":"2019-10-28T15:32:35","modified_gmt":"2019-10-28T19:32:35","slug":"continental-divide-4-oceans","status":"publish","type":"post","link":"https:\/\/blogs.mathworks.com\/steve\/2009\/05\/15\/continental-divide-4-oceans\/","title":{"rendered":"Locating the US continental divide, part 4 – Ocean masks"},"content":{"rendered":"
\r\n

Welcome to part 4 of my series<\/a> on computing the US continental divide. Previously we looked at data import and display, the watershed transform and label\r\n matrices, and different kinds of minima.\r\n <\/p>\r\n

Today we're going to use some binary image manipulation to create an \"ocean mask.\" Let's start by taking a fresh look at some\r\n of the pixel values in the DEM (digital elevation model) we've been working with.\r\n <\/p>

s = load('continental_divide'<\/span>);\r\ndem = s.dem_cropped;\r\ndisplay_range = [-500 3000];\r\nimshow(dem, display_range, 'InitialMagnification'<\/span>, 'fit'<\/span>)<\/pre> 
What are the pixel values associated with the ocean?  You might expect them to be 0. After all, the ocean is at sea level!\r\n       But a few values at the upper-left corner of the DEM are not 0:\r\n   <\/p>
dem(1:5, 1:5)<\/pre>
\r\nans =\r\n\r\n   -500   -500   -500   -500   -500\r\n   -500   -500   -500   -500   -500\r\n   -500   -500   -500   -500   -500\r\n   -500   -500   -500   -500   -500\r\n   -500   -500   -500   -500   -500\r\n\r\n<\/pre>
Are the ocean pixels equal to -500 all the way up to the coast? When I want to inspect pixel values closely, I often use the\r\n      Pixel Region Tool<\/a>. Here's a screen shot of the tool showing pixel values along the coast of California:\r\n   <\/p>\r\n   
 <\/p>\r\n   
So the value -500 seems to be used as an ocean marker value. But I prefer not to guess. After some browsing of the online report<\/a> about this data set, I found this sentence in Section 11<\/a>: \"Every tile contains values of -500 for oceans, with no values between -500 and the minimum value for land noted here.\"\r\n   <\/p>\r\n   
We can use a simple relational operator to form an ocean mask image.<\/p>
ocean_mask = dem == -500;\r\nimshow(ocean_mask, 'InitialMagnification'<\/span>, 'fit'<\/span>)<\/pre> 
I expected ocean_mask<\/tt> to have just two connected components, corresponding to the Pacific and Atlantic Oceans. But I was surprised:\r\n   <\/p>
ocean_mask_labeled = bwlabel(ocean_mask);\r\nmax(ocean_mask_labeled(:))<\/pre>
\r\nans =\r\n\r\n   835\r\n\r\n<\/pre>
Oops. More than 800 labeled regions, not two. Let's zoom in the coast of the Gulf of Mexico to see what's going on.<\/p>
axis([3700 4300 2500 2800])<\/pre> 
You can see that there are \"ocean\" pixels that are completely disconnected from the rest of the oceans.  How can we identify\r\n      just the connected components of ocean pixels on the left and right side of the DEM?  One method is to use bwselect<\/tt>.  This function can find all foreground pixels that are \"connected\" to a given set of seed pixels. Here I'll use it to find\r\n      all the \"ocean\" pixels connected to the upper left and upper right components.\r\n   <\/p>
[M, N] = size(dem);\r\nocean_mask = bwselect(ocean_mask, [1 N], [1 1]);<\/pre>
Here's the same zoomed-in view along the Gulf of Mexico.<\/p>
imshow(ocean_mask, 'InitialMagnification'<\/span>, 'fit'<\/span>)\r\naxis([3800 4400 2550 2750])<\/pre> 
You can see that the interior \"ocean pixels\" have been removed. Let's add this image to our continental divide MAT-file.<\/p>
save continental_divide<\/span> ocean_mask<\/span> -append<\/span><\/pre>
Next time we'll use a technique called minima imposition<\/i> to modify the DEM so that it contains only two regional minima, one for each ocean.\r\n   <\/p>\r\n   
About this Series<\/i><\/p>\r\n   
This series<\/a> explores the problem of computing the location of the continental divide for the United States.  The divide separates the\r\n      Atlantic and Pacific Ocean catchment basins for the North American continent.\r\n   <\/p>\r\n   
As an algorithm development problem, computing the divide lets us explore aspects of data import and visualization, manipulating\r\n      binary image masks, label matrices, regional minima, and the watershed transform.\r\n   <\/p>\r\n   
\r\n