Find the surface area of an object using stereo vision.

Hello,

I want to find out the surface area of tree (only green part) using stereo vision. I know stereo vision and using the same I got the (X,Y,Z) values of tree, (of course one side only) (file is attached). I used delaunay function to calculate area but the problem is that I have some negative values of X and Y which cause error in surface area. How I can get the correct value of surface area using stereo vision?

It will be good if I can calculate volume also.

Thank You.

[Adam Danz]

If the negative values are the only issue, why not offset all of the coordinates so the minimum values are (0,0,0)?

[John D'Errico]

Adam is of course correct. While negative values in your data should never compromise an area computation in correctly written code, if your code is unable to handle them, the simple solution is to just translate ALL of your data so that none are negative. Find the minimum (most negative) value in each dimension, then add a constant to all coordinates in that dimension.

This operation does not change the area. If it does in any significant way beyond floating point trash in the results, then again your code has a problem.

[Naseeb Gill]

Same I was thinking. But unable to do so. What I want to do is remove negative value and add that much value in positive value so that area reain same.

Say for example I have A = [-2,-7,0.5,14,3] by doing offset It should be convert into B =[0,0,2.5,21,3,].

I want to know how to do so? Also a case may be arises when negative values may be more than positive values like C = [-1,-9,-5,14,3]. How to solve this?

[Adam Danz]

"Say for example I have A = [-2,-7,0.5,14,3] by doing offset It should be convert into B ="[0,0,2.5,21,3,]. "

Why would A be converted to that? A would be offset by it's minimum value which is -7.

A = [-2,-7,0.5,14,3]

A = A - min(A)
A =
            5            0          7.5           21           10

For C

            
C = [-1,-9,-5,14,3]

C - min(C)
ans =
     8     0     4    23    12

Note that even if A or C do not have negative values, applying the same proceedure (A - min(A)) does not affect the area.

Answer 1

[KSSV]

After using delaunayTriangulation, you have vertices and nodal connectivity in hand. Then you can use this function to get the area of each triangular element........sum the total areas, you will get the total sutgace area.

Have a look on function triarea from the package.

https://in.mathworks.com/matlabcentral/fileexchange/25555-mesh2d-delaunay-based-unstructured-mesh-generation

[Naseeb Gill]

triarea is in mesh-cost. will you please elaborate me how to use this? Thanks.

[Naseeb Gill]

I used below code: with x,y and z values I got answer = 99.32 but when I insert any negative value anywhere x = [0;-2;-4;6;8;9;0;2;4;6;8;9;0;2;4;6;8;9]; answer cames more than 99. How to solve this?

x = [0;2;4;6;8;9;0;2;4;6;8;9;0;2;4;6;8;9];
y = [0;0;0;0;0;0;5;5;5;5;5;5;10;10;10;10;10;10];
z = [22;21.3;21;21.4;22.2;23.5;22;21.3;21;21.4;22.2;23.5;22;21.3;21;21.4;22.2;23.5];
tri = delaunay(x,y);
P = [x,y,z];
v1 = P(tri(:,2), :) - P(tri(:,1), :);
v2 = P(tri(:,3), :) - P(tri(:,2), :);
cp = 0.5*cross(v1,v2);
surfaceArea = sum(sqrt(dot(cp, cp, 2)))

[KSSV]

The order of the vertices is important..maybe your messing with that. Did you try with triarea?

[Naseeb Gill]

Yes I tried with triarea, it shows error.

Error using triarea (line 29)
Incorrect input dimensions.

Mine pp array is 18x3 and tt array is 24x4.

Also, code I wrote in above comment is same as code written in triarea. Some of its part is:

%--------------------------------------- compute signed area
    ev12 = pp(tt(:,2),:)-pp(tt(:,1),:) ;
    ev13 = pp(tt(:,3),:)-pp(tt(:,1),:) ;

    switch (size(pp,2))
        case +2
           
        area = ev12(:,1).*ev13(:,2) ...
             - ev12(:,2).*ev13(:,1) ;
        area = 0.5 * area;
            
        case +3
            
        avec = cross(ev12,ev13);
        area = sqrt(sum(avec.^2,2)) ;
        area = 0.5 * area;    
        
        otherwise
        error('Unsupported dimension.') ;
    end

After observing both codes, there is no difference. I'm not changing order of vertices, I'm using them as it as.