Sobel and Laplace edge detection in SciPy - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of Sobel filter on a simple image

What is the output array after applying the Sobel filter on the given 3x3 image using scipy.ndimage.sobel with axis=0?

SciPy

import numpy as np
from scipy import ndimage

image = np.array([[10, 10, 10],
                  [20, 20, 20],
                  [30, 30, 30]])
sobel_vertical = ndimage.sobel(image, axis=0)
print(sobel_vertical)

A

[[10 10 10]
 [0 0 0]
 [-10 -10 -10]]

B

[[10 10 10]
 [20 20 20]
 [20 20 20]]

C

[[10 10 10]
 [20 20 20]
 [0 0 0]]

D

[[10 10 10]
 [20 20 20]
 [10 10 10]]

💡 Hint

Think about how the Sobel filter calculates the gradient along the vertical axis.

The Sobel filter computes the gradient by subtracting pixel values above and below. The output shows positive values where intensity increases downward and negative where it decreases.

❓ data_output

intermediate

Number of edges detected by Laplace filter

After applying the Laplace filter on the given 5x5 image, how many pixels have a non-zero value indicating edges?

SciPy

import numpy as np
from scipy import ndimage

image = np.array([
    [10, 10, 10, 10, 10],
    [10, 50, 50, 50, 10],
    [10, 50, 100, 50, 10],
    [10, 50, 50, 50, 10],
    [10, 10, 10, 10, 10]
])
laplace = ndimage.laplace(image)
non_zero_count = np.count_nonzero(laplace)
print(non_zero_count)

A 8

B 12

C 20

D 16

💡 Hint

Edges are where the Laplace output is not zero, usually around intensity changes.

The Laplace filter highlights regions with rapid intensity changes. Counting non-zero pixels gives the number of detected edges.

🔧 Debug

advanced

Identify the error in Sobel filter application

What error will this code raise when applying the Sobel filter incorrectly?

SciPy

import numpy as np
from scipy import ndimage

image = np.array([[1, 2], [3, 4]])
sobel = ndimage.sobel(image, axis=2)
print(sobel)

A ValueError: input array must be 3D

B TypeError: unsupported operand type(s) for +: 'int' and 'str'

C IndexError: axis 2 is out of bounds for array of dimension 2

D No error, prints the Sobel filtered array

💡 Hint

Check the axis parameter against the image array dimensions.

The image is 2D, so axis=2 is invalid and causes an IndexError.

❓ visualization

advanced

Visual difference between Sobel and Laplace filters

Which option best describes the visual difference when applying Sobel and Laplace filters on the same image?

A Sobel and Laplace produce identical edge maps.

B Sobel produces a blurred image; Laplace produces a sharpened image.

C Sobel detects color changes; Laplace detects brightness changes only.

D Sobel highlights edges in one direction; Laplace highlights edges regardless of direction with zero-crossings.

💡 Hint

Think about directional vs. non-directional edge detection.

Sobel calculates gradient in a specific direction, while Laplace detects edges by second derivatives and zero-crossings.

🚀 Application

expert

Combining Sobel and Laplace for edge enhancement

Given an image, which code snippet correctly combines Sobel filters on both axes and the Laplace filter to create an enhanced edge image?

A

sobel_x = ndimage.sobel(image, axis=0)
sobel_y = ndimage.sobel(image, axis=1)
laplace = ndimage.laplace(image)
enhanced = np.sqrt(sobel_x**2 + sobel_y**2) + laplace

B

sobel_x = ndimage.sobel(image, axis=0)
sobel_y = ndimage.sobel(image, axis=1)
laplace = ndimage.laplace(image)
enhanced = sobel_x + sobel_y + laplace

C

sobel = ndimage.sobel(image)
laplace = ndimage.laplace(image)
enhanced = sobel * laplace

D

sobel_x = ndimage.sobel(image, axis=0)
laplace = ndimage.laplace(image)
enhanced = sobel_x - laplace

💡 Hint

Combine gradients from both directions using Pythagorean theorem before adding Laplace.

The combined magnitude of Sobel x and y gradients gives edge strength; adding Laplace enhances edges further.