Padding Borders of an Image with Zeros or Any Constant Value, Using MATLAB

Why Padding of borders is required While working with an Image?

When you apply a filter to pixels on the borders of an image, some of the elements of the masking block or filter kernel may not overlap actual image pixels. For example, if the kernel is 3-by-3 and you are computing the result for a pixel on the top row of the image, some of the elements of the kernel are outside the border of the image. For example see the below image,

 (Image illustration Courtesy, CUVT Prague)

This technique is also used by the software like MS Word, PhotoShop when we want to add a border in the image.

Padding zeros in the border of an Image: MATLAB Implementation

Suppose you are having any image;

Original Grayscale image ready to be padded with zero or any other constant value

MATLAB Code for padding Zeros on the boundary of an image:

clc
clear all
close all
i=imread('xyz.jpg'); %read the desired image
%i=rgb2gray(i); %Uncomment this to Convert the image to grayscale if original one is RGB
imshow(i) %Show the original image before preprocessing
p=input('Input no. of rows/columns to pad:');  

%Ask the user to get the number of rows/columns to be padded
pad=uint8(zeros(size(i)+2*p));  

%generate an initial zero matrix with the dimension added with 2*p more.
for x=1:size(i,1)
            for y=1:size(i,2)
                pad(x+p,y+p)=i(x,y); 

%accessing the x+p & y+p location and storing  x,y location pixel into that

            end
end
figure,imshow(pad) %Show the padded zero image

Result of above code to pad zero in an image:

Input no. of rows/columns to pad:6

Zeros padded grayscale image, Using MATLAB

Original Dimension if the image: 640x640

Resultant Dimension if the image: 652x652

A part of the resultant Image with a corner looks like: 

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    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      0        0        0        0        0
0    0    0    0    0    0      0        0        0        0        0
0    0    0    0    0    0    217    217    217    217    217
0    0    0    0    0    0    217    217    217    216    216
0    0    0    0    0    0    217    217    217    216    216
0    0    0    0    0    0    218    218    217    217    216
0    0    0    0    0    0    219    218    218    217    216
0    0    0    0    0    0    217    217    217    216    216
0    0    0    0    0    0    214    214    214    214    214
0    0    0    0    0    0    212    212    212    212    212
 

MATLAB Code for padding Any Constant Number on the boundary of an image:

%With a little modification of the above code we can achieve that

%The new code will be like

clc
clear all
close all
i=imread('xyz.jpg'); %read the desired image
%i=rgb2gray(i); %Uncomment this to Convert the image to grayscale if original one is RGB
imshow(i) %Show the original image before preprocessing
p=input('Input no. of rows/columns to pad:');

%Ask the user to get the number of rows/columns to be padded

v=input('The value of the constant to be padded:');
pad=uint8(ones(size(i)+2*p)*v);  

%generate an initial zero matrix with the dimension added with 2*p more.
for x=1:size(i,1)
            for y=1:size(i,2)
                pad(x+p,y+p)=i(x,y); 

%accessing the x+p & y+p location and storing  x,y location pixel into that

            end
end
figure,imshow(pad) %Show the constant padded image

Now after an execution of the code:

Input no. of rows/columns to pad:4
Value of the costant? 4

 Our Padded Image matrix will look like:

4    4    4    4      4        4        4        4        4
4    4    4    4      4        4        4        4        4
4    4    4    4      4        4        4        4        4
4    4    4    4      4        4        4        4        4
4    4    4    4    217    217    217    217    217
4    4    4    4    217    217    217    216    216
4    4    4    4    217    217    217    216    216
4    4    4    4    218    218    217    217    216
4    4    4    4    219    218    218    217    216

Author: Vibhutesh Kumar Singh

An active & prominent author at Digital iVision Labs! Specializing in MATLAB, C++, Arduino, OpenCV, NI Labview, Web Designing & other Electronics stuffs! Just started M.Tech. From IIIT Delhi, looking for excellent PhD Opportunities with prominent researchers. Drop a mail: vibhutesh[at]gmail.com or Follow him at....

Facebook

Twitter

Google+

Linkedin