Don't understand the following script for image evaluation

Hey guys,

I have a script for image evaluation (for an image like in the attachment) and unfortunately I don't understand one part. The reading of the image I understand. I don't understand the middle part with "%%%%FINDING Lengths in TOP ROW" (I have marked the part boldly).

What exactly happens there?

Any help would be greatly appreciated!

________________________________________________________________________

,%referncesize=input('Please Enter Refernce size in nm/Pixel: ');

close all

clear

clc

referncesize=690;

%Opening image

[imagefiles ,pathname] = uigetfile({'*.bmp'},'MultiSelect', 'on'); inputimage=strcat(pathname, imagefiles);

files = string(inputimage);

imfiles = string(imagefiles);

for k = 1:length(files)

try

disp(files(k))

I = imread(files(k));

J=rgb2gray(I);

J = imadjust(J);

J = medfilt2(J);

level = graythresh(J);

BW=imbinarize(J,level);

BW = medfilt2(BW);

se = strel('disk',0,0);

erodedBW = imerode(BW,se);

%%%%% CHANGE HERE FOR OTHER RESULTS: USUALLY: [...]('disk',10,0)

se = strel('disk',0,0);

erodedBW = imdilate(erodedBW,se);

figure

imshow(erodedBW)

%%%%FINDING Lengths in TOP ROW

[row,column]=size(erodedBW);

startingpointr1=[];

endingpointr1=[];

startingflag=0;

tempflag=0;

for i=1:column

if erodedBW(1,i)==1

tempflag=tempflag+1;

if tempflag>5

if startingflag==0

startingpointr1=[startingpointr1 i];

startingflag=1;

end

end

else

if tempflag>5

startingflag=0;

endingpointr1=[endingpointr1 i];

end

tempflag=0;

end

end

%%%% FINDING LENGTH IN MIDDLE

startingpointr2=[];

endingpointr2=[];

startingflag=0;

tempflag=0;

for i=1:column

if erodedBW(row/2,i)==1

tempflag=tempflag+1;

if tempflag>5

if startingflag==0

startingpointr2=[startingpointr2 i];

startingflag=1;

end

end

else

if tempflag>5

startingflag=0;

endingpointr2=[endingpointr2 i];

end

tempflag=0;

end

end

%%%% FINDING LENGTH IN Bottom

startingpointr3=[];

endingpointr3=[];

startingflag=0;

tempflag=0;

for i=1:column

if erodedBW(row,i)==1

tempflag=tempflag+1;

if tempflag>5

if startingflag==0

startingpointr3=[startingpointr3 i];

startingflag=1;

end

end

else

if tempflag>5

startingflag=0;

endingpointr3=[endingpointr3 i];

end

tempflag=0;

end

end

%%%%%%%%%%%%%%%FIND COLUMN WISE WIDTH

startingpointc1=[];

endingpointc1=[];

startingflag=0;

tempflag=0;

for i=1:row

if erodedBW(i,1)==1

tempflag=tempflag+1;

if tempflag>5

if startingflag==0

startingpointc1=[startingpointc1 i];

startingflag=1;

end

end

else

if tempflag>5

startingflag=0;

endingpointc1=[endingpointc1 i];

end

tempflag=0;

end

end

%%%%%%%%%COLUMN WISE IN MIDDLE

startingpointc2=[];

endingpointc2=[];

startingflag=0;

tempflag=0;

for i=1:row

if erodedBW(i,(column/2))==1

tempflag=tempflag+1;

if tempflag>5

if startingflag==0

startingpointc2=[startingpointc2 i];

startingflag=1;

end

end

else

if tempflag>5

startingflag=0;

endingpointc2=[endingpointc2 i];

end

tempflag=0;

end

end

%%%%%%%%%COLUMN WISE IN END

startingpointc3=[];

endingpointc3=[];

startingflag=0;

tempflag=0;

for i=1:row

if erodedBW(i,(column))==1

tempflag=tempflag+1;

if tempflag>5

if startingflag==0

startingpointc3=[startingpointc3 i];

startingflag=1;

end

end

else

if tempflag>5

startingflag=0;

endingpointc3=[endingpointc3 i];

end

tempflag=0;

end

end

%startingpointr1

%endingpointr1

r1c1=endingpointr1(1)-startingpointr1(1);

r1c2=endingpointr1(2)-startingpointr1(2);

%startingpointr2

%endingpointr2

r2c1=endingpointr2(1)-startingpointr2(1);

r2c2=endingpointr2(2)-startingpointr2(2);

%startingpointr3

%endingpointr3

r3c1=endingpointr3(1)-startingpointr3(1);

r3c2=endingpointr3(2)-startingpointr3(2);

%startingpointc1

%endingpointc1

r11c1=endingpointc1(1)-startingpointc1(1);

r22c1=endingpointc1(2)-startingpointc1(2);

%startingpointc2

%endingpointc2

r11c2=endingpointc2(1)-startingpointc2(1);

r22c2=endingpointc2(2)-startingpointc2(2);

%startingpointc3

%endingpointc3

r11c3=endingpointc3(1)-startingpointc3(1);

r22c3=endingpointc3(2)-startingpointc3(2);

T1 = {'Position' 'Value (nm)';

'r1c1' r1c1*referncesize;

'r1c2' r1c2*referncesize;

'r2c1' r2c1*referncesize;

'r2c2' r2c2*referncesize;

'r3c1' r3c1*referncesize;

'r3c2' r3c2*referncesize;

'r11c1' r11c1*referncesize;

'r22c1' r22c1*referncesize;

'r11c2' r11c2*referncesize;

'r22c2' r22c2*referncesize;

'r11c3' r11c3*referncesize;

'r22c3' r22c3*referncesize;

};

T = cell2table(T1);

writetable(T,'data.xlsx','Sheet',imfiles(k),'WriteVariableNames',false)

catch

T1 = {'Position' 'Value (nm)';};

T = cell2table(T1);

writetable(T,'data.xlsx','Sheet',imfiles(k),'WriteVariableNames',false)

end

end

Answer 1

[Simran]

Hi @Meike,

I see you want to understand what the “Finding lengths in TOP ROW” part of the code you mentioned. The explaination goes as below:

1.)

[row, column] = size(erodedBW); 
startingpointr1 = []; 
endingpointr1 = []; 
startingflag = 0; 
tempflag = 0;  

Here “startingpointr1” and “endingpointr1” are initialised as emplty arrays. These will store the starting and ending column indices of detected segments in the top row. Then “startingflag” and “emptyflag” are intialised to 0. These flags help in detecting the start and end of a continuous segment of pixels.

2.)

% Loop through each column in the first row 
for i = 1:column 
if erodedBW(1, i) == 1 % Check if the pixel is part of a segment 
tempflag = tempflag + 1; % Increment the temporary counter 
if tempflag > 5 % Check if the segment length is significant 
if startingflag == 0 % If a new segment is detected 
startingpointr1 = [startingpointr1 i]; % Record the starting point 
startingflag = 1; % Set the flag indicating a segment has started 
end 
end 
else 
if tempflag > 5 % If the segment ends and was significant 
startingflag = 0; % Reset the starting flag 
endingpointr1 = [endingpointr1 i]; % Record the ending point 
end 
tempflag = 0; % Reset the temporary counter 
end 
end 

Here the code is divided into two segments- “Detection of Segments” and “End of Segment”.

“Detection of Segments”-

1.) Here If a pixel is part of a segment (erodedBW(1, i) == 1), ”tempflag” is incremented. If ”tempflag” exceeds 5, it then indicates there is a continuous segment of at least 6 pixels, hence suggesting a significant structure.

2.) If the “startingflag” is 0, it then means the start of a new segment is detected, and the current index ”i” is added to ”startingpointr1”. ”startingflag” is then set to 1 to mark that a segment has started.

“End of Segment”

1.) If a pixel is not part of a segment (erodedBW(1, i) == 0), and tempflag is greater than 5, it then marks the end of a segment. The current index ”i” is added to ”endingpointr1”, and ”startingflag” is reset to 0.

2.) “tempflag” is reset to 0 whenever a non-segment pixel is encountered.

This process effectively identifies the start and end of segments in the top row of the binary image, allowing the calculation of lengths of these segments.

You can refer these documentation links for more details on the functions used in the script:

imread - https://www.mathworks.com/help/releases/R2021a/matlab/ref/imread.html

imadjust - https://www.mathworks.com/help/releases/R2021a/images/ref/imadjust.html

imbinarize - https://www.mathworks.com/help/releases/R2021a/images/ref/imbinarize.html

imerode - https://www.mathworks.com/help/releases/R2021a/images/ref/imerode.html

imdilate - https://www.mathworks.com/help/releases/R2021a/images/ref/imdilate.html