Fitting the data so that one a single continuous line is achived

Hi,

I have following line that I want to smooth out so that only one point will be there in one column:

mat file for the data is also attached. I tried to look for curve fitting and interpolation etc but could not comprehend how exactly tackle this.

Answer 1

[KSSV]

I = imread('Annotation 2019-09-04 112619.png') ; 
I1 = imcrop(I) ; % crop to remove white borders present around the image 
I1 = rgb2gray(I1) ; 

[y,x] = find(I1) ; 
% Do fitting 
p = polyfit(x,y,2) ; 
xi = min(x):max(x) ;
yi = polyval(p,xi) ; 

imshow(I1)
hold on
plot(xi,yi,'r')

[waqas]

This is sacrifcing the shape too much which would later cause problems for me. Any other suggestions on how to remove same values column wise (unique function didnt work)?

[KSSV]

Try using smooth

imshow(I1)
hold on
plot(x,smooth(y),'r')

[waqas]

It is still giving almost the same fit. I reckon if some sort of trignomatric fit can give better conservation of orginal randomness of line. Line is basically depicting a crack and showing it as a smooth curve would be a simplified case which would cause problems in later stages of analysis.