Rank: 351570 based on 0 downloads (last 30 days) and 0 file submitted
photo

Pietro Pala

E-mail

Personal Profile:
Professional Interests:
computer vision

 

Watch this Author's files

 

Comments and Ratings by Pietro
Updated File Comments Rating
14 Jun 2012 Iterative Closest Point An implementation of various ICP (iterative closest point) features. Author: Jakob Wilm

I'm trying to test the accuracy of this routine on a known point distribution: model points are the [x y z] coordinates of a Gaussian function; data points are obtained by translating by T the (x,y) coordinates of the model points (data = model + repmat([1.5 1.5 0], size(model,1), 1)). Even for small values of T the routine cannot converge to the optimum. Here is the code used for the test:

k=0;
for x=1:50
for y=1:50
z(x,y) = exp(-(single(x) -25.0).^2 ./400.0 -(single(y)-25.0).^2 ./400.0);
k = k+1;
model(k,:)=[x y z(x,y)];
end
end

surf(double([1:50]), double([1:50]), double(z));

data = model + repmat([15.5 15.5 0], size(model,1), 1);

[TR, TT, ER] = icp(double(model'), double(data'), 20)

Any suggestion? Thx

Contact us