How to add a border around a NumPy array?

Adding a border around a NumPy array is useful in image processing, data visualization, and scientific computing. NumPy provides several methods to add borders: zero padding, constant padding, and concatenation.

In this article, we will explore three different techniques to add borders around NumPy arrays with practical examples.

Method 1: Zero Padding with np.pad()

The simplest method uses np.pad() with mode='constant' (default padding value is 0) ?

import numpy as np

# Create a sample 2D array
original_array = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])

# Add border of width 1 with zeros
bordered_array = np.pad(original_array, pad_width=1, mode='constant')

print("Original Array:")
print(original_array)
print("\nWith Zero Border:")
print(bordered_array)

Original Array:
[[1 2 3]
 [4 5 6]
 [7 8 9]]

With Zero Border:
[[0 0 0 0 0]
 [0 1 2 3 0]
 [0 4 5 6 0]
 [0 7 8 9 0]
 [0 0 0 0 0]]

Method 2: Constant Value Padding

Use constant_values parameter to specify a custom border value ?

import numpy as np

original_array = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])

# Add border with constant value 99
bordered_array = np.pad(original_array, pad_width=2, mode='constant', constant_values=99)

print("Original Array:")
print(original_array)
print("\nWith Constant Border (99):")
print(bordered_array)

Original Array:
[[1 2 3]
 [4 5 6]
 [7 8 9]]

With Constant Border (99):
[[99 99 99 99 99 99 99]
 [99 99 99 99 99 99 99]
 [99 99  1  2  3 99 99]
 [99 99  4  5  6 99 99]
 [99 99  7  8  9 99 99]
 [99 99 99 99 99 99 99]
 [99 99 99 99 99 99 99]]

Method 3: Manual Concatenation

Create border arrays manually and concatenate them with the original array ?

import numpy as np

def add_border_concatenation(arr, border_width, border_value=0):
    height, width = arr.shape
    
    # Create border arrays
    top_border = np.full((border_width, width), border_value)
    bottom_border = np.full((border_width, width), border_value)
    
    # Add top and bottom borders
    temp_array = np.concatenate([top_border, arr, bottom_border], axis=0)
    
    # Create left and right borders
    new_height = temp_array.shape[0]
    left_border = np.full((new_height, border_width), border_value)
    right_border = np.full((new_height, border_width), border_value)
    
    # Add left and right borders
    final_array = np.concatenate([left_border, temp_array, right_border], axis=1)
    
    return final_array

original_array = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])
bordered_array = add_border_concatenation(original_array, border_width=1, border_value=5)

print("Original Array:")
print(original_array)
print("\nWith Concatenation Border:")
print(bordered_array)

Original Array:
[[1 2 3]
 [4 5 6]
 [7 8 9]]

With Concatenation Border:
[[5 5 5 5 5]
 [5 1 2 3 5]
 [5 4 5 6 5]
 [5 7 8 9 5]
 [5 5 5 5 5]]

Advanced Padding Options

NumPy's pad() function offers several padding modes for different use cases ?

import numpy as np

original_array = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])

# Reflect padding (mirror values)
reflect_border = np.pad(original_array, pad_width=1, mode='reflect')

# Edge padding (repeat edge values)
edge_border = np.pad(original_array, pad_width=1, mode='edge')

print("Reflect Padding:")
print(reflect_border)
print("\nEdge Padding:")
print(edge_border)

Reflect Padding:
[[5 4 5 6 5]
 [2 1 2 3 2]
 [5 4 5 6 5]
 [8 7 8 9 8]
 [5 4 5 6 5]]

Edge Padding:
[[1 1 2 3 3]
 [1 1 2 3 3]
 [4 4 5 6 6]
 [7 7 8 9 9]
 [7 7 8 9 9]]

Comparison

Conclusion

Use np.pad() for most border-adding tasks as it's efficient and flexible. Choose concatenation when you need complex custom border patterns that np.pad() cannot achieve.