Erosion and Dilation of images using OpenCV in Python (original) (raw)
Last Updated : 05 Aug, 2025
Morphological operations modify images based on the structure and arrangement of pixels. They apply kernel to an input image for changing its features depending on the arrangement of neighboring pixels. Morphological operations like erosion and dilation are techniques in image processing, especially for binary or grayscale images. They help in analyzing shapes, cleaning noise and refining object boundaries.
Erosion
Erosion in image processing is a morphological operation that shrinks and thins the boundaries of objects in an image by removing pixels on object edges, effectively making objects smaller and removing small white noise.
Purpose
- Shrinks or erodes the boundaries of foreground objects (usually white pixels).
- Removes fine white noise and separates objects that are touching.
How It Works
- A kernel (usually a 3×3, 5×5 or 7×7 matrix of ones) slides across the image.
- A pixel remains white (1) only if all pixels under the kernel are white; otherwise, it becomes black (0).
- This process reduces object size and erodes edges.
Dilation
Dilation is a morphological operation that expands the boundaries of objects in an image by adding pixels to object edges making objects appear larger and filling in small gaps or holes.
**Purpose:
- Expands the boundaries of the foreground objects.
- Accentuates or enlarges features and fills small gaps.
**How It Works:
- The kernel is similarly convolved over the image.
- A pixel is set to white (1) if _at least one of the corresponding pixels under the kernel is white.
- As a result, the white regions grow, merging small holes or joining broken parts together.
Implementation of Erosion and Dilation
Let's implement Erosion and Dilation with OpenCV in Python,
**Step 1: Import Libraries
We will import the necessary libraries,
- **cv2****:** OpenCV library for image processing.
- **numpy****:** For numerical operations and to create kernels.
- **matplotlib.pyplot****:** To display images in notebooks.
**Step 2: Load Input Image and Define the Structuring Elements(Kernel)
The kernel defines the neighborhood for the operation. Common choices are rectangles or disks.
Used image can be downloaded from here.
Python `
img = cv2.imread('input.png', 0) plt.imshow(img, cmap='gray') plt.title("Original Image") plt.axis('off') plt.show() kernel = np.ones((5, 5), np.uint8)
`
**Output:
Original
**Step 3: Apply Erosion
Erosion works by sliding the kernel across the image. A pixel remains white (255) only if all pixels under the kernel are white, otherwise, it becomes black (0). This reduces object boundaries and removes small white noise.
Python `
img_erosion = cv2.erode(img, kernel, iterations=1)
plt.imshow(img_erosion, cmap='gray') plt.title("After Erosion") plt.axis('off') plt.show()
`
**Output:
After Erosion
**Step 4: Apply Dilation
Dilation slides the kernel across the image and a pixel becomes white if at least one pixel under the kernel is white. This thickens white regions or objects and fills small holes.
Python `
img_dilation = cv2.dilate(img, kernel, iterations=1)
plt.imshow(img_dilation, cmap='gray') plt.title("After Dilation") plt.axis('off') plt.show()
`
**Output:
After Dilation
Applications
**Erosion
- Removing isolated white noise from an image.
- Separating objects that are joined or touching.
- Finding object boundaries by shrinking object size.
**Dilation
- Filling small holes or gaps in objects.
- Joining broken or disconnected parts of the same object.
- Used after erosion (as part of "opening" operation) to restore object size while keeping noise removed.
Erosion and dilation are fundamental morphological operations in image processing that allow us to refine, clean and manipulate shapes within images. By using simple structuring elements, these techniques help remove noise, separate or connect objects and enhance image features making them essential tools for effective pre-processing and analysis in computer vision tasks with OpenCV and Python.