Assignment.py

"""
🔷 Contours & Shape Detection App

\======================================

This assignment helps users explore the concept of **contour detection** and **shape approximation** using OpenCV.
It demonstrates how to detect objects, approximate their geometry, and annotate their shape on an image.

---

📋 **Instructions**:

1. Start the program or notebook.

2. ## Choose one of the following shape creation methods:

   **\[1] Draw Basic Shapes (Circle, Rectangle, etc.)**

   * Programmatically create shapes (e.g., using `cv2.circle`, `cv2.rectangle`)
   * Shapes must be filled or outlined with white (value = 255) on a black background

   **\[2] Use a Preloaded or Uploaded Image**

   * Upload an image with clearly defined shapes
   * Convert to grayscale and apply thresholding before contour detection

---

3. ## Perform Contour Detection:

   * Convert the image to grayscale (if not already)
   * Apply **thresholding** (e.g., `cv2.threshold`) to get a binary image
   * Use `cv2.findContours()` to extract contours

---

4. ## Shape Approximation & Classification:

   * Iterate through contours
   * Use `cv2.approxPolyDP()` to approximate each contour's shape
   * Count the number of corners:
     - **3 points** → Triangle
     - **4 points** → Rectangle or Square
     - **5 points** → Pentagon
     - **>5 points** → Circle (or other curved shapes)

---

5. ## Draw and Annotate Results:

   * Use `cv2.drawContours()` to draw detected shapes in color
   * Use `cv2.putText()` to label each shape with its name above the contour
   * You may also print coordinates or area (optional)

---

📦 **Optional Additions**:

- Detect nested contours using `cv2.RETR_TREE` and draw hierarchy
- Add area/centroid calculations using `cv2.contourArea()` and `cv2.moments()`
- Annotate number of shapes detected

---

💡 **Notes**:

* Ensure the input image is **binary or high contrast** for effective contour detection
* Use shape approximation threshold like `0.01 * cv2.arcLength(...)` for precision
* All contours should be labeled clearly with shape name and drawn in color for clarity

---

🎓 **Learning Objectives**:

✅ Understand how contours represent shape boundaries in images
✅ Learn to use polygonal approximation to detect shape types
✅ Practice using OpenCV functions: `findContours`, `approxPolyDP`, `drawContours`, `putText`
✅ Improve image analysis and object recognition skills with OpenCV

---

🔧 **Dependencies**:

* `OpenCV (cv2)`
* `NumPy`
* `matplotlib` *(for image display)*
* `PIL (optional)` *(for custom image processing or saving)*
"""

import cv2
import numpy as np
import matplotlib.pyplot as plt


def create_test_image():
    """Create an image with different shapes on a black background."""
    img = np.zeros((400, 400), dtype=np.uint8)
    cv2.rectangle(img, (50, 50), (150, 150), 255, -1)  # Square
    cv2.circle(img, (300, 100), 50, 255, -1)  # Circle
    cv2.polylines(
        img, [np.array([[200, 200], [250, 300], [150, 300]])], True, 255, -1
    )  # Triangle
    cv2.polylines(
        img,
        [np.array([[270, 220], [320, 270], [300, 320], [250, 320], [230, 270]])],
        True,
        255,
        -1,
    )  # Pentagon
    return img


def detect_shapes(img):
    contours, _ = cv2.findContours(img, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
    annotated = cv2.cvtColor(img, cv2.COLOR_GRAY2BGR)

    for cnt in contours:
        epsilon = 0.01 * cv2.arcLength(cnt, True)
        approx = cv2.approxPolyDP(cnt, epsilon, True)
        x, y, w, h = cv2.boundingRect(approx)

        # Shape classification
        corners = len(approx)
        if corners == 3:
            shape = "Triangle"
        elif corners == 4:
            shape = "Rectangle"
        elif corners == 5:
            shape = "Pentagon"
        else:
            shape = "Circle"

        cv2.drawContours(annotated, [approx], 0, (0, 255, 0), 2)
        cv2.putText(
            annotated, shape, (x, y - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (255, 0, 0), 2
        )

    return annotated


if __name__ == "__main__":
    img = create_test_image()
    _, thresh = cv2.threshold(img, 127, 255, cv2.THRESH_BINARY)
    result = detect_shapes(thresh)

    plt.figure(figsize=(10, 5))
    plt.subplot(1, 2, 1)
    plt.title("Input Image")
    plt.imshow(img, cmap="gray")

    plt.subplot(1, 2, 2)
    plt.title("Detected Shapes")
    plt.imshow(cv2.cvtColor(result, cv2.COLOR_BGR2RGB))
    plt.axis("off")
    plt.show()