Customed distance function Haversine

For a proyect I want to use pdist2 (Pairwise distance between two sets of observations) but I need an specific function.

I have two data sets of different sizes, one of which is a nx2 matrix of latitude, longitude.

I'm trying to calculate Haversine distance but I don't know how to apply the funtion in this case. Would you help correcting my function? Thank you in advance.

axa = pdist2(latlon1(:, 1:2), latlon2(:, 1:2), @haversine);

function [dist] = haversine(lat1,lon1,lat2,lon2)
   dist = 2 * 6372.8 * asin(sqrt(sind((lat2-lat1)/2)^2 + cosd(lat1) * cosd(lat2) * sind((lon2 - lon1)/2)^2));
end

[SALAH ALRABEEI]

pdist2 has only three metric distances ( 'seuclidean', 'minkowski', or 'mahalanobis').

So since your lon and lat are not of the same length; you can create grid

[x,y]=meshgrid(lon,lat); 

Answer 1

[Scott MacKenzie]

You need to re-work your haversine function, as per the requirements for distance functions used with pdist2. A custom distance function for pdist2 needs two input arguments and the first must specify just a single observation:

axa = pdist2(latlon1(1, 1:2), latlon2(:, 1:2), @haversine); % repeat for other values in latlon1

function [dist] = haversine(ZI, ZJ)
   lat1 = ZI(1,1);
   lon1 = ZI(1,2);
   lat2 = ZJ(:,1);
   lon2 = ZJ(:,2);
   dist = 2 * 6372.8 * asin(sqrt(sind((lat2-lat1)/2)^2 + cosd(lat1) * cosd(lat2) * sind((lon2 - lon1)/2)^2));
end

[Miguel Angel Sanchez Aportela]

Thank you so much, it worked really well!

[Scott MacKenzie]

@Miguel Angel Sanchez Aportela You're welcome. Glad to help. Good luck with your research.