Book Image

Raspberry Pi Computer Vision Programming - Second Edition

By : Ashwin Pajankar
5 (1)
Book Image

Raspberry Pi Computer Vision Programming - Second Edition

5 (1)
By: Ashwin Pajankar

Overview of this book

Raspberry Pi is one of the popular single-board computers of our generation. All the major image processing and computer vision algorithms and operations can be implemented easily with OpenCV on Raspberry Pi. This updated second edition is packed with cutting-edge examples and new topics, and covers the latest versions of key technologies such as Python 3, Raspberry Pi, and OpenCV. This book will equip you with the skills required to successfully design and implement your own OpenCV, Raspberry Pi, and Python-based computer vision projects. At the start, you'll learn the basics of Python 3, and the fundamentals of single-board computers and NumPy. Next, you'll discover how to install OpenCV 4 for Python 3 on Raspberry Pi, before covering major techniques and algorithms in image processing, manipulation, and computer vision. By working through the steps in each chapter, you'll understand essential OpenCV features. Later sections will take you through creating graphical user interface (GUI) apps with GPIO and OpenCV. You'll also learn to use the new computer vision library, Mahotas, to perform various image processing operations. Finally, you'll explore the Jupyter Notebook and how to set up a Windows computer and Ubuntu for computer vision. By the end of this book, you'll be able to confidently build and deploy computer vision apps.
Table of Contents (15 chapters)

Performing transformation operations on images

In this section, we will learn how to perform various mathematical transformation operations on images with OpenCV and Python 3.

Scaling

Scaling means resizing an image. It is a geometric operation. OpenCV offers a function, cv2.resize(), for performing this operation. It accepts an image, a method for the interpolation of pixels, and the scaling factor as arguments and returns a scaled image. The following methods are used for the interpolation of the pixels in the output:

  • cv2.INTER_LANCZOS4: This deals with the Lanczos interpolation method over a neighborhood of 8x8 pixels.
  • cv2.INTER_CUBIC: This deals with the bicubic interpolation method over a neighborhood of 4x4 pixels and is preferred for performing the zooming operation on an image.
  • cv2.INTER_AREA: This means resampling using pixel area relation. This is preferred for performing the shrinking operation on an image.
  • cv2.INTER_NEAREST: This means the method...