Image Processing Exercise

What you'll learn

• How to resize images using different scaling factors
• Converting images between color spaces (Gray, HSV, RGB)
• Applying edge detection with Canny Edge
• Creating blur effects with Gaussian filters
• Working with OpenCV functions in Python

Introduction

This exercise will guide you through various image processing techniques using OpenCV in Python. You'll practice image resizing and conversion to different color spaces.

Access the exercise materials here: Drive Link

Preparing Folders and Files

⏱️ Time needed: 5 minutes

Instructions: Follow the steps below

1. Download the C14-Image-Processing-template folder from Google Drive into your Program G folder
2. Extract the folder to your working directory
3. Open the extracted folder in VSCode
4. For those without VSCode, please complete your exercise by following the guidelines here

:bulb: Make sure you have OpenCV installed by running pip install opencv-python in your terminal before starting!

:computer: Read, Think, Code

⏱️ Time needed: 30 minutes

Question One: Rescale image

⏱️ Estimated time: 10 minutes

In resize_image.py file:

1. Resize the image to scale 0.2
2. Create variables for scale, width, height and dim
3. Apply the cv.resize() syntax and replace the original image
4. Show the resized image using cv.imshow() tip Remember that scale 0.2 means the image will be 20% of its original size. Calculate the new dimensions using:

• width = int(img.shape[1] * scale)
• height = int(img.shape[0] * scale)

Question 2: Convert image

⏱️ Estimated time: 20 minutes

In convert_image.py file:

1. Convert image to Gray color

   • Apply the cv.cvtColor() syntax with cv.COLOR_BGR2GRAY and store it as gray
   • Show the gray image using cv.imshow('Gray', gray)

2. Convert image to HSV color

   • Apply the cv.cvtColor() syntax with cv.COLOR_BGR2HSV and store it as hsv
   • Show the HSV image using cv.imshow('HSV', hsv)

3. Convert image to RGB color

   • Apply the cv.cvtColor() syntax with cv.COLOR_BGR2RGB and store it as rgb
   • Show the RGB image using cv.imshow('RGB', rgb)

4. Convert image to Canny Edge

   • Apply the cv.Canny() syntax with threshold values (100, 200) and store it as canny
   • Show the Canny Edge image using cv.imshow('Canny', canny)

5. Convert image to Blur

   • Apply the cv.GaussianBlur() syntax with kernel size (7,7) and store as blur
   • Show the blur image using cv.imshow('Blur', blur)

:bulb: For Canny Edge detection, experiment with different threshold values to see how they affect edge detection sensitivity!

The final output will be shown as follows:

:art: Try it Yourself

⏱️ Time needed: 15 minutes

Create a main.py file and practice with your own image:

1. Find and load an image
   • Use any image from your computer or download one
   • Load it using cv.imread('your_image.jpg')
2. Apply resize operations
   • Try different scale factors (0.5, 0.75, 1.5)
   • Experiment with different interpolation methods
3. Apply multiple conversions
   • Combine different effects (e.g., convert to gray then apply blur)
   • Save your favorite results using cv.imwrite() tip Try chaining effects! For example, convert to grayscale first, then apply Canny Edge detection for cleaner edge results.

Exercise Submission

:warning: Important! Make sure to submit your completed work before the deadline!

Please submit your work through this link: Exercise Submission Form

What to submit:

• Your completed resize_image.py file
• Your completed convert_image.py file
• Your main.py file with custom experiments
• Screenshots of your output images

Troubleshooting

Common Issues:

Image not loading?

• Check the file path is correct
• Ensure the image file exists in the same directory
• Try using absolute path instead of relative path

OpenCV not found?

• Install with: pip install opencv-python
• For Anaconda users: conda install opencv

Window not showing?

• Add cv.waitKey(0) after cv.imshow()
• Add cv.destroyAllWindows() at the end of your script

Colors look wrong?

• Remember OpenCV loads images in BGR format, not RGB
• Use cv.COLOR_BGR2RGB when needed