Morphological operations (erosion, dilation, opening, closing) in Computer Vision - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to apply erosion on the image using OpenCV.

Computer Vision

eroded_image = cv2.[1](image, kernel, iterations=1)

Drag options to blanks, or click blank then click option'

Aerode

Bdilate

Cblur

Dthreshold

Attempts:

3 left

💡 Hint

Common Mistakes

Using dilate instead of erode will expand white regions instead of shrinking them.

Using blur or threshold functions do not perform morphological erosion.

The erode function shrinks bright regions in the image, which is the erosion operation.

2fill in blank

medium

Complete the code to create a 3x3 rectangular structuring element for morphological operations.

Computer Vision

kernel = cv2.getStructuringElement(cv2.MORPH_[1], (3, 3))

Drag options to blanks, or click blank then click option'

ARECT

BCROSS

CELLIPSE

DSQUARE

Attempts:

3 left

💡 Hint

Common Mistakes

Choosing ELLIPSE or CROSS will create different shaped kernels, not rectangular.

SQUARE is not a valid OpenCV structuring element shape.

The RECT shape creates a rectangular kernel, commonly used in morphological operations.

3fill in blank

hard

Fix the error in the code to perform morphological opening on the image.

Computer Vision

opened_image = cv2.morphologyEx(image, cv2.MORPH_[1], kernel)

Drag options to blanks, or click blank then click option'

ADILATE

BCLOSE

CERODE

DOPEN

Attempts:

3 left

💡 Hint

Common Mistakes

Using CLOSE applies dilation then erosion, which is the opposite of opening.

Using ERODE or DILATE alone does not perform opening.

MORPH_OPEN applies erosion followed by dilation, which is the opening operation.

4fill in blank

hard

Fill both blanks to perform morphological closing on the image with a 5x5 elliptical kernel.

Computer Vision

kernel = cv2.getStructuringElement(cv2.MORPH_[1], (5, 5))
closed_image = cv2.morphologyEx(image, cv2.MORPH_[2], kernel)

Drag options to blanks, or click blank then click option'

AELLIPSE

BOPEN

CCLOSE

DRECT

Attempts:

3 left

💡 Hint

Common Mistakes

Using RECT instead of ELLIPSE changes the kernel shape.

Using OPEN instead of CLOSE applies the wrong morphological operation.

Using ELLIPSE creates an elliptical kernel, and MORPH_CLOSE applies dilation followed by erosion, which is closing.

5fill in blank

hard

Fill all three blanks to create a dilation operation with a 7x7 cross-shaped kernel and 2 iterations.

Computer Vision

kernel = cv2.getStructuringElement(cv2.MORPH_[1], (7, 7))
dilated_image = cv2.[2](image, kernel, iterations=[3])

Drag options to blanks, or click blank then click option'

ACROSS

Bdilate

C2

Derode

Attempts:

3 left

💡 Hint

Common Mistakes

Using ERODE instead of dilate will shrink regions instead of expanding.

Using RECT or ELLIPSE changes the kernel shape.

Setting iterations to 1 or missing it changes the effect strength.

The CROSS kernel shape is used, dilate applies dilation, and 2 sets the number of iterations.

Practice

1. What does the erosion operation do to the white parts of a binary image?

easy

A. It grows the white parts by adding pixels at the edges.

B. It removes noise by smoothing the edges.

C. It fills small holes inside the white parts.

D. It shrinks the white parts by removing pixels at the edges.

Solution

1. Step 1: Understand erosion effect on white pixels

   Erosion removes pixels from the boundaries of white regions, making them smaller.

2. Step 2: Compare with other operations

   Dilation grows white parts, opening removes noise, closing fills holes, so erosion must shrink white parts.

3. Final Answer:

   It shrinks the white parts by removing pixels at the edges. -> Option D

4. Quick Check:

   Erosion = Shrinks white parts [OK]

Hint: Erosion shrinks white areas by cutting edges [OK]

Common Mistakes:

• Confusing erosion with dilation
• Thinking erosion fills holes
• Mixing erosion with noise removal

2. Which of the following is the correct syntax to perform dilation using OpenCV in Python on an image img with a 3x3 kernel?

easy

A. cv2.erode(img, np.ones((3,3), np.uint8))

B. cv2.dilate(img, np.ones((3,3), np.uint8))

C. cv2.dilate(img, (3,3))

D. cv2.dilate(img, kernel=3)

Solution

1. Step 1: Recall dilation syntax in OpenCV

   Dilation requires the image and a structuring element (kernel), usually created with np.ones of desired size and type.

2. Step 2: Check options for correct usage

   cv2.dilate(img, np.ones((3,3), np.uint8)) uses cv2.dilate with a 3x3 kernel created by np.ones and correct dtype, which is valid syntax.

3. Final Answer:

   cv2.dilate(img, np.ones((3,3), np.uint8)) -> Option B

4. Quick Check:

   Dilation syntax = cv2.dilate(image, kernel) [OK]

Hint: Use np.ones((3,3), np.uint8) as kernel for dilation [OK]

Common Mistakes:

• Using erode instead of dilate
• Passing kernel size tuple directly
• Using wrong kernel datatype

3. Given the following Python code using OpenCV:

import cv2
import numpy as np
img = np.array([[0,0,0,0,0],
                [0,255,255,255,0],
                [0,255,0,255,0],
                [0,255,255,255,0],
                [0,0,0,0,0]], dtype=np.uint8)
kernel = np.ones((3,3), np.uint8)
eroded = cv2.erode(img, kernel)
print(eroded)

What will be the printed output?

medium

A. [[ 0 0 0 0 0]
[ 0 255 0 255 0]
[ 0 0 0 0 0]
[ 0 255 0 255 0]
[ 0 0 0 0 0]]

B. [[ 0 0 0 0 0]
[ 0 255 255 255 0]
[ 0 255 255 255 0]
[ 0 255 255 255 0]
[ 0 0 0 0 0]]

C. [[ 0 0 0 0 0]
[ 0 0 0 0 0]
[ 0 0 0 0 0]
[ 0 0 0 0 0]
[ 0 0 0 0 0]]

D. [[255 255 255 255 255]
[255 255 255 255 255]
[255 255 255 255 255]
[255 255 255 255 255]
[255 255 255 255 255]]

Solution

1. Step 1: Understand erosion on the given image

   Erosion removes pixels at edges of white regions. The center pixel (0) surrounded by 255s will cause erosion to shrink the white area.

2. Step 2: Apply 3x3 kernel erosion

   Since the kernel covers neighbors, any pixel with a zero neighbor becomes zero. The white cross shape will erode to a smaller cross with zeros at the center and edges.

3. Final Answer:

   White cross with zeros at center and edges as shown in option A -> Option A

4. Quick Check:

   Erosion shrinks white, so edge pixels vanish but some inner pixels remain [OK]

Hint: Erosion removes edge pixels, shrinking white areas [OK]

Common Mistakes:

• Assuming erosion keeps center pixels
• Confusing erosion with dilation output
• Ignoring zero pixels in kernel neighborhood

4. You wrote this code to perform opening on an image img but it does not remove noise as expected:

kernel = np.ones((3,3), np.uint8)
opened = cv2.dilate(cv2.erode(img, kernel), kernel)

What is the error and how to fix it?

medium

A. Kernel size is too small; increase kernel size to remove noise.

B. The order is reversed; opening is dilation followed by erosion, so swap the calls.

C. Use cv2.morphologyEx with cv2.MORPH_OPEN instead for correct opening.

D. The order is reversed; opening is erosion followed by dilation, so code is correct.

Solution

1. Step 1: Check the definition of opening

   Opening is erosion followed by dilation. The code applies erosion then dilation, which is correct in order.

2. Step 2: Identify practical issue

   Manual calls may work but can be error-prone; using cv2.morphologyEx with MORPH_OPEN is recommended for correct and optimized opening.

3. Final Answer:

   Use cv2.morphologyEx with cv2.MORPH_OPEN instead for correct opening. -> Option C

4. Quick Check:

   Use built-in morphologyEx for opening [OK]

Hint: Use cv2.morphologyEx with MORPH_OPEN for opening [OK]

Common Mistakes:

• Swapping erosion and dilation order
• Not using built-in morphology functions
• Assuming kernel size fixes logic errors

5. You have a noisy binary image with small black holes inside white objects. Which morphological operation should you apply to fill these holes without changing the object shapes much?

hard

A. Closing

B. Dilation

C. Opening

D. Erosion

Solution

1. Step 1: Understand the problem of black holes inside white objects

   Black holes are small dark spots inside white regions that need to be filled.

2. Step 2: Choose operation that fills holes without shrinking objects

   Closing is dilation followed by erosion; it fills small holes and gaps while preserving object shape.

3. Final Answer:

   Closing -> Option A

4. Quick Check:

   Closing fills holes inside white objects [OK]

Hint: Closing fills holes inside white objects [OK]

Common Mistakes:

• Using erosion which shrinks objects
• Using opening which removes noise but not holes
• Confusing dilation alone with closing