Real-Time IoT Imaging with Deep Neural Networks: Using Java on the Raspberry Pi 4

By Nicolas Modrzyk

This book shows you how to build real-time image processing systems all the way through to house automation. Find out how you can develop a system based on small 32-bit ARM processors that gives you complete control through voice commands.

Real-time image processing systems are utilized in a wide variety of applications, such as in traffic monitoring systems, medical image processing, and biometric security systems. In Real-Time IoT Imaging with Deep Neural Networks, you will learn how to make use of the best DNN models to detect object in images using Java and a wrapper for OpenCV. Take a closer look at how Java scripting works on the Raspberry Pi while preparing your Visual Studio code for remote programming. You will also gain insights on image and video scripting. Author Nicolas Modrzyk shows you how to use the Rhasspy voice platform to add a powerful voice assistant and completely run and control your Raspberry Pi from your computer. 

To get your voice intents for house automation ready, you will explore how Java connects to the MQTT and handles parametrized Rhasspy voice commands. With your voice-controlled system ready for operation, you will be able to perform simple tasks such as detecting cats, people, and coffee pots in your selected environment. Privacy and freedom are essential, so priority is given to using open source software and an on-device voice environment where you have full control of your data and video streams. Your voice commands are your own—and just your own.

With recent advancements in the Internet of Things and machine learning, cutting edge image processing systems provide complete process automation. This practical book teaches you to build such a system, giving you complete control with minimal effort.

What You Will Learn:

• Show mastery by creating OpenCV filters
• Execute a YOLO DNN model for image detection
• Apply the best Java scriptingon Raspberry Pi 4
• Prepare your setup for real-time remote programming
• Use the Rhasspy voice platform for handling voice commands and enhancing your house automation setup

Who This Book Is For:Engineers, and Hobbyists wanting to use their favorite JVM to run Object Detection and Networks on a Raspberry Pi

• Language: English
• Publisher: Apress
• Release date: Mar 10, 2020
• ISBN: 9781484257227

Book preview

© Nicolas Modrzyk 2020

N. ModrzykReal-Time IoT Imaging with Deep Neural Networkshttps://doi.org/10.1007/978-1-4842-5722-7_1

1. Getting Started

Nicolas Modrzyk¹ 

(1)

Tokyo, Tokyo, Japan

One of the goals of this book is to get you ready to perform real-time IoT imaging quickly, avoiding a lengthy installation process. Being ready quickly doesn’t mean we are going to take any shortcuts, it means we will get the tooling part out of the way so we can focus on the creation process.

In this chapter, you’ll run your first example.

Visual Studio Code Primer

You can’t knock on opportunity’s door and not be ready.

—Bruno Mars

The playground setup introduced in this book is fairly standard for people who are used to writing code. It also has a little bit of a new kid on the block feeling—only the cool kids use it. The stack , or the environment, we will use consists of the following:

Visual Studio Code, a small but pluggable text editor

The Java Development Kit and its runtime, so we can write Java code and run it

A Java plugin to Visual Studio Code so that the editor understands the Java code and runs it

Basically, that’s all.

It does not really matter whether you’re using Windows, Mac, or Linux. All the different pieces of XX are made to run anywhere. To install Visual Studio Code, you need to head to https://code.visualstudio.com/.

Click the download button on the top-right side of the page. After you click the button, you’ll be presented with options for the different packages, one for each computer platform, as shown in Figure 1-1.

../images/490964_1_En_1_Chapter/490964_1_En_1_Fig1_HTML.jpg

Figure 1-1

Choosing your download

At the time of writing, the current version is 1.40, but newer versions would only be better.

Running the installer and opening Visual Studio Code for the first time gives you a screen similar to the one in Figure 1-2.

../images/490964_1_En_1_Chapter/490964_1_En_1_Fig2_HTML.jpg

Figure 1-2

Visual Studio Code

The second step is to install Java, if it is not already installed. You can either head to the OpenJDK web site (https://jdk.java.net/) and download a zip file or go to the Oracle web site and download a ready-to-use installer for your machine (see Figure 1-3 and Figure 1-4). Specifically, you can go to https://www.oracle.com/technetwork/java/javase/downloads/index.html.

../images/490964_1_En_1_Chapter/490964_1_En_1_Fig3_HTML.jpg

Figure 1-3

Oracle Java download page

../images/490964_1_En_1_Chapter/490964_1_En_1_Fig4_HTML.jpg

Figure 1-4

Java download link

Let the installer run to the end, and do not stop it even if it tries to open another installer while running. Java, being a development kit, does not come with a fancy application to check that it has been installed properly, so after the Java installation is finished, a quick way to check that things are in place is to open a terminal in Visual Studio Code and check the Java version.

You can press Ctrl+Shift+@ to open a terminal inside Visual Studio Code, or you can open it from the menu, as shown in Figure 1-5.

../images/490964_1_En_1_Chapter/490964_1_En_1_Fig5_HTML.jpg

Figure 1-5

Opening a terminal window from within Visual Studio Code

This will pop up a small tab, usually at the bottom of the editor, as shown in Figure 1-6.

../images/490964_1_En_1_Chapter/490964_1_En_1_Fig6_HTML.jpg

Figure 1-6

The Terminal tab

Then inside the terminal, type the following command:

java -version

Figure 1-7 shows the expected version output if you have installed the OpenJDK