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)

Perspective transformation of images

In the mathematical operation of perspective transformation, a set of four points in the input image is mapped to a set of four points in the output image. The criteria for selecting the set of four points in the input and the output image is that any three points (in the input and the output image) must not be in the same line. Like affine mathematical transformation, in perspective transformation, the straight lines in the input images remain straight. However, there is no guarantee that the parallel lines in the input image remain parallel in the output image.

One of the most prominent real-life examples of this mathematical operation is the zoom and the angled zoom functions in image editing and viewing software tools. The amount of zoom and angle of zooming depend on the matrix of the transformation that is computed by the two sets of points that we discussed earlier. OpenCV provides the cv2.getPerspectiveTransform() function, which accepts...