Image stitching for mapping in Drone Programming - Mini Project: Build & Apply

Image Stitching for Mapping

📖 Scenario: You are programming a drone to create a map by stitching together images it takes while flying over a field. Each image has a unique ID and a position coordinate. Your task is to combine these images into a single stitched map based on their positions.

🎯 Goal: Build a simple program that takes a dictionary of images with their positions, sets a stitching threshold, selects images close enough to stitch, and outputs the list of images that will be stitched together.

📋 What You'll Learn

Create a dictionary called images with image IDs as keys and their (x, y) positions as values.

Create a variable called stitch_threshold to set the maximum distance between images to be stitched.

Use a for loop with variables img1 and pos1 to iterate over images.items().

Inside the loop, use another for loop with variables img2 and pos2 to iterate over images.items().

Calculate the distance between pos1 and pos2 and if it is less than or equal to stitch_threshold, add the pair (img1, img2) to a list called stitch_pairs.

Print the stitch_pairs list.

💡 Why This Matters

🌍 Real World

Drones capture many images of land or buildings. Stitching these images creates a complete map or model.

💼 Career

Understanding image stitching helps in drone programming, mapping, surveying, and geographic information systems (GIS).

1

Create the images dictionary

Create a dictionary called images with these exact entries: 'img1': (0, 0), 'img2': (3, 4), 'img3': (6, 8), 'img4': (10, 10).

Drone Programming

# Create the images dictionary with exact keys and position tuples
# Your code here

Hint

Use curly braces {} to create a dictionary. Each key is a string like 'img1' and each value is a tuple with two numbers.

2

Set the stitching threshold

Create a variable called stitch_threshold and set it to 5.

Drone Programming

images = {'img1': (0, 0), 'img2': (3, 4), 'img3': (6, 8), 'img4': (10, 10)}
# Set stitch_threshold to 5
# Your code here

Hint

Just write stitch_threshold = 5 on a new line.

3

Find image pairs to stitch

Create an empty list called stitch_pairs. Use a for loop with variables img1 and pos1 to iterate over images.items(). Inside it, use another for loop with variables img2 and pos2 to iterate over images.items(). Calculate the distance between pos1 and pos2 using the formula ((pos1[0] - pos2[0])**2 + (pos1[1] - pos2[1])**2)**0.5. If the distance is less than or equal to stitch_threshold and img1 is not equal to img2, add the tuple (img1, img2) to stitch_pairs.

Drone Programming

images = {'img1': (0, 0), 'img2': (3, 4), 'img3': (6, 8), 'img4': (10, 10)}
stitch_threshold = 5
# Create stitch_pairs list and add pairs of images within threshold
# Your code here

Hint

Remember to check that img1 is not the same as img2 to avoid pairing an image with itself.

4

Print the stitching pairs

Write print(stitch_pairs) to display the list of image pairs that will be stitched together.

Drone Programming

images = {'img1': (0, 0), 'img2': (3, 4), 'img3': (6, 8), 'img4': (10, 10)}
stitch_threshold = 5
stitch_pairs = []
for img1, pos1 in images.items():
    for img2, pos2 in images.items():
        distance = ((pos1[0] - pos2[0])**2 + (pos1[1] - pos2[1])**2)**0.5
        if distance <= stitch_threshold and img1 != img2:
            stitch_pairs.append((img1, img2))
# Print the stitch_pairs list
# Your code here

Hint

Just write print(stitch_pairs) to see the pairs.

Practice

1. What is the main purpose of image stitching in drone programming?

easy

A. To combine multiple drone photos into one large map

B. To control the drone's flight path

C. To improve the drone's battery life

D. To send live video feed to the operator

Solution

1. Step 1: Understand image stitching concept

   Image stitching is about joining many photos to create a bigger picture.

2. Step 2: Identify the main use in drone programming

   It helps create large maps from many drone images for better area coverage.

3. Final Answer:

   To combine multiple drone photos into one large map -> Option A

4. Quick Check:

   Image stitching = combine photos [OK]

Hint: Image stitching means joining photos into one big map [OK]

Common Mistakes:

• Confusing image stitching with drone control
• Thinking it improves battery life
• Mixing it with live video streaming

2. Which of the following is the correct way to start stitching images side by side in a drone program?

easy

A. stitch(images, method='side_by_side')

B. stitch(images, method='fly_path')

C. stitch(images, method='battery_save')

D. stitch(images, method='live_feed')

Solution

1. Step 1: Identify stitching method names

   Side by side stitching joins images horizontally or vertically.

2. Step 2: Match method to stitching task

   Only 'side_by_side' matches the stitching method; others relate to flight or battery.

3. Final Answer:

   stitch(images, method='side_by_side') -> Option A

4. Quick Check:

   Method 'side_by_side' = stitching images side by side [OK]

Hint: Use method='side_by_side' to join images simply [OK]

Common Mistakes:

• Using unrelated methods like 'fly_path' or 'battery_save'
• Confusing stitching method with drone flight commands
• Misspelling method names

3. Given the code snippet:

images = ['img1.jpg', 'img2.jpg', 'img3.jpg']
stitched_map = stitch(images, method='overlap')
print(len(stitched_map))

What will be the output if stitched_map is a combined image with 3 parts?

medium

A. 1

B. 3

C. 0

D. Error

Solution

1. Step 1: Understand the stitch function output

   Stitching combines images into one big image, so the result is a single image object.

2. Step 2: Check what len() returns on the stitched image

   Most image objects do not support len(), so calling len() will cause an error.

3. Final Answer:

   Error -> Option D

4. Quick Check:

   len() on image object causes error [OK]

Hint: len() on stitched image causes error [OK]

Common Mistakes:

• Assuming length equals number of input images
• Expecting len() to work on image object
• Confusing list length with image object length

4. The following code is meant to stitch images with overlap detection but throws an error:

images = ['a.jpg', 'b.jpg']
stitched = stitch(images, method='overlap')
stitched.show()

What is the likely cause of the error?

medium

A. The method 'overlap' is not supported

B. The show() method does not exist for stitched images

C. The stitch function requires images to be loaded, not file names

D. The images list is empty

Solution

1. Step 1: Check input type for stitch function

   Stitch usually needs image data, not just file names as strings.

2. Step 2: Identify error cause

   Passing file names without loading images causes errors when stitching or showing.

3. Final Answer:

   The stitch function requires images to be loaded, not file names -> Option C

4. Quick Check:

   Input must be loaded images, not file names [OK]

Hint: Load images before stitching, not just file names [OK]

Common Mistakes:

• Assuming method 'overlap' is invalid
• Thinking show() method is missing
• Ignoring that images list is not empty

5. You want to create a stitched map from drone images but only include images that have at least 30% overlap with the previous one. Which approach best applies this condition in your stitching code?

hard

A. Stitch all images first, then crop the map to overlapping areas

B. Filter images by overlap percentage before stitching using a function

C. Use method='side_by_side' to ignore overlap

D. Manually stitch images without any overlap checks

Solution

1. Step 1: Understand the overlap condition

   Only images with at least 30% overlap should be included in stitching.

2. Step 2: Choose the best way to apply this condition

   Filtering images before stitching ensures only suitable images are combined.

3. Final Answer:

   Filter images by overlap percentage before stitching using a function -> Option B

4. Quick Check:

   Filter first, then stitch = correct approach [OK]

Hint: Filter images by overlap before stitching [OK]

Common Mistakes:

• Stitching all images then cropping
• Ignoring overlap and stitching side by side
• Manual stitching without overlap checks