Introduction to IoT with Machine Learning and Image Processing using Raspberry Pi

Shrirang Ambaji Kulkarni, Varadraj P. Gurupur, Steven L. Fernandes, Varadrah P. Gurupur

• Publisher: Chapman and Hall/CRC
• Year: 2020
• Print ISBN: 9781138543522
• eBook ISBN: 9781351006644
• Edition: 1
• Topic: Computer Science
• Subtopic: Computer Science General
• Index: Computer Science

Machine Learning a branch of Artificial Intelligence is influencing the society, industry and academia at large. The adaptability of Python programming language to Machine Learning has increased its popularity further. Another technology on the horizon is Internet of Things (IoT). The present book tries to address IoT, Python and Machine Learning along with a small introduction to Image Processing. If you are a novice programmer or have just started exploring IoT or Machine Learning with Python, then this book is for you.

Features:

• Raspberry Pi as IoT is described along with the procedure for installation and configuration.
• A simple introduction to Python Programming Language along with its popular library packages like NumPy, Pandas, SciPy and Matplotlib are dealt in an exhaustive manner along with relevant examples.
• Machine Learning along with Python Scikit-Learn library is explained to audience with an emphasis on supervised learning and classification.
• Image processing on IoT is introduced to the audience who love to apply Machine Learning algorithms to Images
• The book follows hands-on approach and provide a huge collection of Python programs.

Chapter 1

Introduction

In the last three decades, technology has changed the way we live our lives. This happens to be a universally accepted truth. In this book, we attempt to illustrate the use of a new technology that has changed the world of communication, computing and computing education. Interestingly, the name of this technology coincides with that of a famous dessert named Raspberry Pie. Raspberry Pi is a single board computer developed in the United Kingdom. Incidentally, the organization that started this is known as the Raspberry Pi Foundation. This was first launched in the year 2012.

By the year 2015, Raspberry Pi had gained wide-scale popularity. One of the key features of this technology was its size and adaptability. In some strange ways, Raspberry Pi was the panacea the world of technology was dreaming of. To build on this idea of a tiny computer, one must consider the fact that more than a decade ago you would have computer programs running on desktops wired to electromechanical systems that would use them. The advent of the Raspberry Pi enables the community of technology developers to innovate handheld devices than can encompass the power of regular computers. Our intention in writing this book is twofold: a) we want to provide a basic understanding of how a Raspberry Pi can be used for simple applications, and b) we want to provide fundamental information on how a Raspberry Pi can be used to advance innovations in machine learning and image recognition.

While providing fundamental information on Raspberry Pi, we will help students comprehend the necessary information required to develop applications and devices. These devices and applications can have a wide range of applications such as networking devices, using devices and applications related to privacy and security, creating medical applications and developing sensors. The authors believe that this range of applications will help the scientific community adapt Raspberry Pi for use in some of their projects and thereby enhance its usability.

Additionally, the authors are attempting to present information on how Raspberry Pi can successfully implement machine learning and image processing. The information presented in this book will help the community of researchers synthesize new scientific methods, algorithms, devices and other forms of technology into a wide range of application domains. The authors envision Raspberry Pi advancing the development of science and technology across various user domains such as medicine, security, communication, and the military. Last but not least, it is our attempt to have this book used as study material for courses teaching students about using Raspberry Pi. It is our understanding that Raspberry Pi can be used extensively in healthcare information systems.

In this book, the authors present a transformative interdisciplinary perspective of Raspberry Pi usage. This idea is based on the philosophy of transformative transdisciplinary perspective presented by the founding fathers of Society for Design and Process Science [1]. This philosophy is based on bringing about a positive transformative change to one domain by making changes or improvements in another. This philosophy is applicable to Raspberry Pi because rapid improvements in Raspberry Pi and its applicability will facilitate and create a rostrum for improvements in all the domains in which this device can be used. One example will be the implementation of machine learning algorithms that can aid image processing. In this book, the authors have dwelled on some of the fundamental programming concepts that are needed for image analysis and machine learning. This is another very important reason this book is needed for its targeted audience.

While we engage ourselves in recognizing the importance of the Raspberry Pi, it is also important to note that Raspberry Pi has competitors or alternatives available in the market. A brief summary of these alternatives is listed in Table 1.1.

Table  1.1  A List of Raspberry Pi Alternatives Available as of 2019 [2]

Raspberry Pi Model	Brief Description
Odroid XU4	Compatible with a few prominent versions of Linux operating system and comparable to Raspberry Pi 3
UDOO Bolt	Works well with desktop applications
ASUS Tinker Board	Works with versions of Linux operating system and Chrome
LattePanda Alpha	Unlike many alternatives; supports windows 10 and uses an Intel Core M3 processor
Banana Pi M64	Works with a wider range of operating systems
RockPro 64	A powerful 64 bit CPU
BeagleBone Black	Developed by Texas Instruments in collaboration with Newark element 14
Libre Computer AML-S905X-CC Le Potato	Suitable for image processing
MinnowBoard Turbot	Its small size and affordability make it a viable competitor for the Raspberry Pi.
Odroid H2	Suitable for game streaming and video applications
Arduino	Useful for applications involving robotics

This book is divided into five chapters. The first chapter provides some introductory material on Raspberry Pi and a brief introduction to the authors. The second chapter explains different types of Raspberry Pis and provides the reader with important information on the slots and other necessary elements of the Raspberry Pi. An explanation on these elements or parts is essential from a usability perspective. It also provides instructions on installing the operating system Raspbian [3] on Raspberry Pi. Additionally, it provides some information on peripheral devices and the ways in which they can be used with the Raspberry Pi.

The importance of using Raspbian dwells in the idea that installing and using this operating system is a necessary step in the synthesis of systems and applications. Raspbian is a freely available operating system for the Raspberry Pi. It is important to note that Raspbian is not affiliated with the Raspberry Pi Foundation. Raspbian is a community-funded development effort as was the case in the early days of the Linux operating system. Interestingly, Raspbian is a Linux distribution built with the Linux operating system. Raspbian was first released in the year 2012.

The third chapter explains the elements of Python programming [4] that are essential with respect to using Raspberry Pi. Here the authors dwell on esoteric programming details that are needed for successful Raspberry Pi application. These details include matrix operations, Cholesky decomposition and modifying data frames. It is important to note that data manipulation and analysis have become critical in today’s world of artificial intelligence and big data. Raspberry Pi provides the much-needed hardware for manipulating big data. The authors understand this need and provide some basic information on the development of systems for data manipulation using the Raspberry Pi. Chapter 3 contains necessary program snippets with illustrations of expected outcomes. One of the critical objectives of this illustration is to help the reader with machine learning application development using the Raspberry Pi. These applications can be applied to smart phones, wearable devices, farming devices and other IoT applications.

Chapter 4 dwells on details with respect to programming machine-learning systems. The authors explore required approaches for splitting data sets into test and training sets essential for machine learning algorithm implementations. We use health informatics examples to explain how machine learning can be implemented using the Raspberry Pi. This chapter provides the reader with necessary information on synthesizing expert systems, developing knowledge banks and performing knowledge curation.

Machine learning implementation is explained with several examples, programing snippets and figures associated with the outcome of their implementation. Implementation of machine learning is important in many domains including health informatics, where machine learning has been extensively used to predict healthcare outcomes and diagnoses and for genetics and bioinformatics. Machine learning and big data analytics have been helpful in predicting the early onset of diseases. Additionally, the possibility of the development of a disease or disorder later in life can be predicted with the application of bioinformatics based on the information derived from human genomes. The integration of Raspberry Pi into portable biomedical devices enhances the possibility of computation.

Finally, chapter 5 deals with image processing using the Raspberry Pi. Here the authors provide many program snippets and associated outputs, thereby delineating details that are necessary for the reader to use the device to perform image-processing algorithms. In recent times, it has been observed that image processing and machine learning associate well. Thus, the authors provide information on machine learning and image processing in this book. With the advent of the Raspberry Pi, the size of the computational device for image processing is reduced while at the same time it improved the available computability. This aids the development and synthesis of real-time sensors in need of advanced processing.

The authors have extensive experience in research and development of science and technology. Dr. Shrirang Kulkarni has a Doctoral degree in Computer Engineering; he started his career in research by exploring and developing technologies for ad hoc wireless networks. He further identified other areas of science and technology in which his fundamental approaches could be used, thereby expanding on the transdisciplinary and transformative nature of his research. He now works on research projects that involve using the Raspberry Pi for developing machine learning techniques for healthcare. Additionally, Dr. Kulkarni has written several books that have attained national attention in India. It is also important to mention that he has been involved in teaching engineering students for the last 18 years.

Dr. Steven Fernandes is a well-accomplished researcher in the area of Internet of Things (IoT), machine learning, image processing and the design and development of intelligent systems. He has published more than 40 articles in reputable journals. Additionally, he has served as a guest editor for several special issues. Dr. Fernandes has been involved in teaching engineering students in India and the United States for about a decade.

Finally, Dr. Varadraj Prabhu Gurupur is currently serving as an Associate Professor with the Department of Health Management and Informatics at the University of Central Florida. Dr. Gurupur has more than 100 publications, including a book, chapters, journal articles, abstracts, conference papers and published reviews. He has worked on several projects funded by agencies such as the National Science Foundation and the National Institutes of Health. He has been actively involved with the Institute of Electrical and Electronic Engineers for over a decade. Additionally, he has been honored with several national, regional and state level awards in the United States for his accomplishments in the area of Health Informatics. Dr. Gurupur received his Master’s in Computer Science in the year 2005 and Doctor of Philosophy in Computer engineering in 2010 from the University of Alabama at Birmingham.

REFERENCES

[1]    Society for Design and Process Science Official Website [Online]. Available: www.sdpsnet.org/sdps/. Accessed: 09/24/2019.

[2]    Electromaker [Online]. Available: www.electromaker.io/blog/article/10-best-raspberry-pi-alternatives. Accessed: 09/24/2019.

[3]    Raspbian [Online]. Available: www.raspbian.org/RaspbianImages

[4]    S.A. Kulkarni, Problem Solving and Python Programming, YesDee Publishers, Chennai, India, 2017.

Chapter 2

Raspberry Pi Unraveled

2.0 Raspberry Pi

Raspberry Pi is an Advanced RISC Machines (ARM)-processed, credit-card s...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Table of Contents
• Authors
• Chapter 1. Introduction
• Chapter 2. Raspberry Pi Unraveled
• Chapter 3. Python and Its Libraries for Machine Learning
• Chapter 4. Machine Learning
• Chapter 5. Introduction to Image Processing
• Bibliography
• Index