Python Shape Detection ⬛⚫

A beginner-friendly Python project that looks at an image of shapes and automatically outlines and names each one — Triangle, Rectangle, Pentagon, Hexagon, or Circle — using OpenCV.

What it does

The program takes a picture of shapes, cleans it up step by step, finds every shape, counts its corners, and writes the shape's name right in the middle of it.

Step	What happens
1. Open image	Load the picture
2. Greyscale	Remove colour, keep only light and dark
3. Threshold	Make it pure black & white (shapes become white)
4. Morphology	Thicken outlines, close gaps, remove noise
5. Find contours	Trace every shape's outline
6. Name shapes	Count corners → pick the name → label it

How the naming works

The number of corners tells us the shape:

Corners	Shape
3	Triangle
4	Rectangle
5	Pentagon
6	Hexagon
many	Circle

Files in this repo

• code_shape_detection.ipynb — the full commented code (open it in Google Colab)
• shapes.png — a sample input image
• Shape Detection Beginners Guide.pdf — a step-by-step guide explaining every line

How to run

Google Colab (easiest — nothing to install):

1. Open the notebook in Google Colab.
2. Upload shapes.png using the folder icon on the left.
3. Run the cells from top to bottom.

On your own computer:

pip install opencv-python numpy matplotlib pillow

Then run the notebook in Jupyter.

Use your own image

Draw a few shapes on white paper with a thick black marker, take a photo, upload it, and change the file name in the first step:

img = Image.open("your_photo.jpg")

Tuning tips

• Shapes not detected? Change the threshold number 110 (try 90 or 150).
• Outlines too thin? Increase the dilate iterations.
• A shape mis-named? Its outline is probably wobbly — dilate a bit more.

Built with

Python · OpenCV · NumPy · Matplotlib · Pillow