CMPS-4350 MATLAB Image Processing Lab

Work with Image Processing ToolBox on MATLAB.


Setup:

    Download MATLAB 2025 here
    SIGN UP USING YOUR CSUB EMAIL SO YOU DONT HAVE TO PAY 

    Goal: Find circle, rectangle, triangle objects in an image
    Goal2: Create your own image for testing




    MATLAB Basic Commands
    
    Original Tutorial: Identifying Round Objects




Find Circular Objects In An Image:

    Setup: Download findShapes.m
    Setup: Download findCircle.m

    "pillsetc.png": Image Provided By MATLAB

    
  
    Step 1: Preprocess the Image 

    1. read image, display, and save to variable    
    
    >> colored_image = imread("pillsetc.png");
    >> displayImage(colored_image, "Original Image")
    

    2. convert rgb image to a binary image
    
    >> grayed = im2gray(colored_image); 
    >> bw = imbinarize(grayed);         
    >> displayImage(bw, "Black and White Image")
    

    3. remove extra pixels and fix empty space
    
    >> minSize = 30;
    >> bw = bwareaopen(bw,minSize); 
    >> displayImage(bw, "Filtered Image")

    >> se = strel("disk",2);
    >> bw = imclose(bw,se); % fill small gaps and holes
    >> displayImage(bw, "Unfilled Fixed Image")

    >> bw = imfill(bw,"holes"); % fill any holes in objects
    >> displayImage(bw, "Filled Fixed Image")
    

    4. find boundaries of objects
    
    >> [B,L] = bwboundaries(bw,"noholes");

    >> labeled_image = label2rgb(L,@jet,[.5 .5 .5]);
    >> displayImage(labeled_image, 'Objects with Boundaries in White')
    >> hold on
    >> for k = 1:length(B)
        boundary = B{k};
        plot(boundary(:,2),boundary(:,1),"w",LineWidth=2)
    >> end
    

    Step 2: Find Circles In Image

    1. get information of the objects
    
    >> stats = regionprops(L, "Circularity", "Centroid");
    

    2.: Set threshold 
    
    >> threshold = 0.94;
    

    3.: Loop through objects to determine shape
    
    >> for k = 1:length(B)
            boundary = B{k};
            circle_value = stats(k).Circularity;

            if circle_value > threshold
                shape = "Circle";
            end

            centroid = stats(k).Centroid;
            text(boundary(1, 2)-35, 
                 boundary(1, 1)+13, 
                 shape, 
                 Color="y", 
                 FontSize=14, 
                 FontWeight="bold")
       end
    >> title("Identified Shapes")
    



Task 1: Make A Custom Image In MATLAB
    
    1. Create a new script named "drawShapeImage"

    2. Create a blank image
        This can be done using zeros function
        it will have 4 parameters
        1. size of the x axis
        2. size of the y axis
        3. rgb color channels (how many channels are there in a rgb image?)
        4. unsigned int typename

    You can create any shape but this is to test the image processing
    The insertShape function will be used to do this

    3. Draw a Rectangle
        call insertShape with a total of 5 parameters and save to img
        1. image you want to insert the shape
        2. filled shape name
        3. dimensions, [x1, y1, width, height]
        4. 'Color' (this one stays the same)
        5. name or rgb value of color you want

    4. Draw a Circle
        call insertShape with a total of 5 parameters and save to img
        1. image you want to insert the shape
        2. filled shape name
        3. dimensions, [x, y, radius]
        4. 'Color' (this one stays the same)
        5. name or rgb of color you want

    5. Draw a Triangle 
        make a variable called "triangle" 
        set it equal to a matrix with each verticies, [x1, y1; x2, y2; x3, y3]
        call insertShape with a total of 5 parameters and save to img
        1. image you want to insert the shape
        2. filled shape name (wont be triangle)
        3. triangle variable
        4. 'Color' (this one stays the same)
        5. name or rgb of color you want

    6. Display And Save The Image
        call imshow function with img as a parameter
        under imshow, put "axis on" to better visual the points
        call imwrite function to save img to a png

     Sample Image Created Using This Method:

    




Task 2: Find Rectangles And Triangles In An Image

    Setup: the "findCircle.m" from above will be used for this section
    
    As all the image setup is done already, you can start from line 56 

    1. Add more parameters to the regionprops function
    
        The regionprops function is a function from the Image Processing Toolbox 
        where it measures several properties from objects in an image
        
        The total parameters for the regionprops function will be 5
        We already know 3 of these parameters: "L", "Circularity", "Centroid"
        
        The 1st parameter will be the one that uses eccentricity for triangle

        The 2nd parameter will be the one that uses solidity for rectangle


    2. Set threshold for each shape

        The threshold for all shapes will be between 0 and 1

        We already have the threshold for a circle being 0.96

        To get the threshold for a triangle and rectangle,
        threshold for triangle that I used "0.6"
        threshold for rectangle that I used "0.9"

    3. Add more logic to the inside of the for loop
        
        We will want to set the values from "stats" saved similar to 
        "circle_value = stats(k).Circularity;"

        After this we will update the if else section to be an if elseif else section 


   >> if circle_value > threshold_circularity
   >>      shape = "Circle";
   >> elseif regionpropvalue < threshold_triangle
   >>     shape = "Triangle";
   >> elseif regionpropvalue > threshold_rectangle 
   >>     shape = "Rectangle";
   >> else
   >>     shape = "Other";
   >> end



    Setup: Download findShapes.m
    Setup: Download findShapes.m
    Setup: Download findShapes.m
    Setup: Download findShapes.m
    Setup: Download findShapes.m
   

    Sample Run Of Finished Lab :