Curve fitting for 2d Array

Hi,

I have to find slope by using curve fitting in this figure. There are 7 wavefronts in this image. What method should I apply? I have tried different techniques like I manually selected 2 points and find slope but that is not appropriate way. I have to utilize all the data points and fit curve to all 7 wavefronts. Sample image is attached here

Answer 1

[KSSV]

If (x,y) are the points of your curve...

m = gradient(y)./gradient(x) ;   % diff(y)./diff(x)

[Leostar90]

What is x and y here? First I have to find Data points of the wavefront and then I have to fit curve.

[KSSV]

In the plot you have a legend, fitted curve. What's that?

[Leostar90]

In first picture, I selected three points manualy. Then using mathematical manipulations I found constants of quadratic equations and then I used "fit" function with "poly2" and then green line was plotted on the wavefront. But in this way I only used 3 points manually. What I want to do is to find many data points or all points of wavefront and then fit a best fitting curve utilizing all data points. Moreover, there are 7 wavefronts and have to do same for all.

[KSSV]

Wavefront has 2D data right? On this use polyfit. Read about polyfit.

[Leostar90]

these are pixel values. I read about polyfit but that need x and y.

[KSSV]

Yes, for the pixel values the positions are (x,y). You need to seek those positions and fit a curve.

[Leostar90]

Yes, I am still there.. How to find those positions? m = gradient(y)./gradient(x) ; will this work? if I pass x and y positions? Size of data is 50*1024 double. so If I pass x 1:1024 and y 1:50, that Way I can find gradient of every position.

Answer 2

[yanqi liu]

yes，sir，may be choose the target line by color and fit them，such as

im = imread('https://www.mathworks.com/matlabcentral/answers/uploaded_files/896015/image.png');
jm = rgb2hsv(im);
v = mat2gray(jm(:,:,3));
bw = ~im2bw(v,0.9);
bw2 = imopen(bw, strel('line', 20, 0));
bw3 = logical(bw-bw2);
bw3 = imopen(imclose(bw3, strel('disk', 5)), strel('square', 3));
bw3 = bwareafilt(bw3, 2);
[r,c] = find(bw3);
p = polyfit(c,r,3);
xc = linspace(min(c),max(c));
yc = polyval(p,xc);
figure; imshow(im)
hold on; plot(xc,yc,'r--','LineWidth',2);

[Leostar90]

Hi, Thank you for your response.

When I try to run this code then I get this error.

" Error using rgb2hsv>parseInputs (line 95)

MAP must be a Mx3 array.

Error in rgb2hsv (line 36)

[r, g, b, isColorMap, isEmptyInput, isThreeChannel] = parseInputs(varargin{:});

Error in region_seg_demo (line 18)

jm = rgb2hsv(im);

I have data in mat file. So I am just loading mat file and and after accessing value from mat file just passing it to your above mentioned code.

Here is the code too.

clear all,clc

dta = load('C:\Users\Awais\Downloads\Imgaes\Imgaes\region_seg\image.mat');
im=dta.L;

jm = rgb2hsv(im); 
v = mat2gray(jm(:,:,3));
bw = ~im2bw(v,0.9);
bw2 = imopen(bw, strel('line', 20, 0));
bw3 = logical(bw-bw2);
bw3 = imopen(imclose(bw3, strel('disk', 5)), strel('square', 3));
bw3 = bwareafilt(bw3, 2);
[r,c] = find(bw3);
p = polyfit(c,r,3);
xc = linspace(min(c),max(c));
yc = polyval(p,xc);
figure; imshow(im)
hold on; plot(xc,yc,'r--','LineWidth',2);

[Leostar90]

did you detect that green line in image or just detected data of that line? I have to apply same curve fitting for all wavefronts in same image? Moreover, I have to find equation of fitted line for finding slope.