Extract Data Points within a Closed Boundary.

I have a large array of geographical data in which the first two columns are latitude and longitude respectively. Remaining columns contain other data like the prominance of a peak located at that particular lat and long. I also have a matrix of lattitude and longitude pairs that define the borders of a particular closed area like a State or Country (thanks to Chad Green's borders.m code in the File Share section).

What I need to do is extract all the data from the large array that are inside or on the border of the particular area I choose. I'm not trying to find the number of points within or on the boundary but looking for a new output matrix which contains the lat & long point along with the other associated column data from the original large array.

I'm guessing Matlab may already contain a function for this operation which is probably faster than any brute force code I would write. I don't have the mapping toobox, just basic Matlab.

Thanks

[jonas]

if you provide some data you are more likely to get solid advice, but check out inpolygon()

[Image Analyst]

How large? Tens of millions of rows? Hundreds of millions of rows? Billions of rows? Anyway I think going down row-by-row using inpolygon() might be the approach to use.

[Shawn]

The input data set can be as large as 25 million rows and 6 to 10 columns. It might be possible to trim it down by some pre-processing, like getting rid of points with a negative latitude if the area of interest is in the northern hemisphere and so forth to focus down to a smaller input matrix.

I'm looking at inploygon now for the first time. At first glance, it looks like it returns a vector containing 1 for an input row that's inside the boundary and 0 for an input row outside the boundary. So for an input matrix of 25 million rows inpolygon would return a vector of equal length.

My typical use for Matlab is feedback control design, so this is new territory for me.

Thanks

Answer 1

[Shawn]

Jonas and Image Analyst

Thanks for the pointer to inpolygon(). That solved my problem and was resonably fast sorting through 7.8 million records, taking only a few seconds.