vgg_H_from_x_lin.m
function H = vgg_H_from_x_lin(xs1,xs2)
% H = vgg_H_from_x_lin(xs1,xs2)
%
% Compute H using linear method (see Hartley & Zisserman Alg 3.2 page 92 in
%                              1st edition, Alg 4.2 page 109 in 2nd edition). 
% Point preconditioning is inside the function.
%
% The format of the xs [p1 p2 p3 ... pn], where each p is a 2 or 3
% element column vector.

[r,c] = size(xs1);

if (size(xs1) ~= size(xs2))
 error ('Input point sets are different sizes!')
end

if (size(xs1,1) == 2)
  xs1 = [xs1 ; ones(1,size(xs1,2))];
  xs2 = [xs2 ; ones(1,size(xs2,2))];
end

% condition points
C1 = vgg_conditioner_from_pts(xs1);
C2 = vgg_conditioner_from_pts(xs2);
xs1 = vgg_condition_2d(xs1,C1);
xs2 = vgg_condition_2d(xs2,C2);

D = [];
ooo  = zeros(1,3);
for k=1:c
  p1 = xs1(:,k);
  p2 = xs2(:,k);
  D = [ D;
    p1'*p2(3) ooo -p1'*p2(1)
    ooo p1'*p2(3) -p1'*p2(2)
   ];
end

% Extract nullspace
[u,s,v] = svd(D, 0); s = diag(s);
 
nullspace_dimension = sum(s < eps * s(1) * 1e3);
if nullspace_dimension > 1
  fprintf('Nullspace is a bit roomy...');
end
 
h = v(:,9);

H = reshape(h,3,3)';

%decondition
H = inv(C2) * H * C1;

H = H/H(3,3);