Simple Image Processing with Python

Python is not a programming language for writing algorithms. It is a dynamic language , secure and high level. If we want to write a complex algorithm we need pointers for efficiency and we will probably use C/C++. Python do have tons of external packages, some of them implemented in C and using a simple interface we can do great (and fast) processing

One popular area in algorithms is Signal processing. If you want to manipulate audio, video, images or any other signal you can use some python packages that implemented in C and you win it all – easy programming and fast execution and response.

Matplotlib and NumPy

Matplotlib package is useful for graph generation and visualisation. We can also use it to read and display images:

Output:

The img object is NumPy array. We can read some properties for example the image shape:

We have 3 dimension array , 768*768 pixels and 4 bytes per pixel: R, G, B, A (alpha).

Lets write a simple function rgb2gray To convert the image to grayscale

you need to calculate for each pixel:

R * 0.299 + G * 0.587 + B * 0.144

We use here NumPy dot product to multiply each pixel value.

Lets display the image again, we need to specify the colormap to be grayscale

The Alpha blending value

The Alpha blending value is not required you can load a image without – means you have only 3 bytes per pixel (RGB). For example:

We need to change the above function to multiple with only 3 values array:

We can make other filters using NumPy functions but we don’t need to “reinvent the wheel”

Using SciPy

Scipy packages built over numpy to extend it with scientific functions

We can use scipy.signal package to run some filters for example:

 

Using SciKit-Image

SciKits are many packages build above SciPy to support different scientific areas like signal processing, RF, and many more. You can see the full list here

We can use scikit-image to manipulate images in a few lines of code. It also contains some sample images built in the library for example:

Now If you want to run a filter you have many functions for example to convert the above image to grayscale :

You can run other filters using the filters package for example sobel filter:

 

As you can see, for practical use, you don’t have to be mathematician , you can use many popular packages to get great results. most of those packages implemented in C so it works fast.

 

 

 

Tagged , ,

2 thoughts on “Simple Image Processing with Python

  1. Thanks so much for sharing such a valuable info. I really appreciate your wonderful explanation. Please keep posting such kind of information on your page. I will certainly dig it and personally suggest to my friends. I’m sure they’ll be benefited from this website.

  2. I often visit your website and have noticed that you don’t update it often. More frequent
    updates will give your website higher authority & rank in google.
    I know that writing content takes a lot of time, but you can always help yourself with miftolo’s
    tools which will shorten the time of creating an article to
    a couple of seconds.

Leave a Reply

Your email address will not be published. Required fields are marked *