In two processed images, one shows 3 centroids and another shows 4 centroids. How to eliminate this extra centroids from centroids array.

I have two images taken from two stereo cameras in which one contain three white objects and another contain four white objects. I processed both images to find out the centroids. Ofcourse first one give three centroids and second give four centroids and I stored these centroids into arrays called 'centroids1' and 'centroids2'. Now, I want two things here. First, if in 'centroids1' first entry is for object 1, then in 'centroids2' also, first entry should be for object 1 only, likewise for all objects. Second, I want to eliminate any extra centroids present in any of 'centroids' array i.e. I want only corresponding centriods in both 'centroids1' and 'centroids2' array.

I'm attaching my images and code here.

Answer 1

[Dima Lisin]

Hi Naseeb,

The triangulate function can return reprojection errors, which can tell you whether or not a match is correct.

[Naseeb Gill]

Thanks Dima, As you suggest triangulate function, using this clue I tried to solve my problem and I came out with this code. It works similar to as I want.

 load('C:\Users\Naseeb\Desktop\centroids1.mat'); 
load('C:\Users\Naseeb\Desktop\centroids2.mat'); 
 B = centroids1; % center of white objects from left cam
A = centroids2;  % center of white objects from right cam
n = size(A,1);
m = size(B,1);
s = n-m;
if s >= 0
   A(m+1:n,:)=[];
else
   B(n+1:m,:)=[];
end
r = reprojectionErrors; % obtain from triangulate function
C = cat(2,B,r(:,1)); % for left image
D = cat(2,A,r(:,1)); % for right image
nA=[];
for j=1:size(C,1)
    if (C(j,3)<3);
        nA=[nA; C(j,:)];
    end
end
nB=[];
for j=1:size(D,1)
    if (D(j,3)<3);
        nB=[nB; D(j,:)];
    end
end

Suggest something if I missed or can be add in this. Also, I'm using triangulate function to find center of white objects,so suggest something regarding how to approach effectively please.

Thanks.