Security Camera with Face Recognition

In Chapter 3, A Social and Personal Digital Photo Frame, you learned how to build a multimedia application, which is a digital photo frame based on the Raspberry Pi. Now you're going to take a major step forward, by building a real-time video processing application, which is a security camera with face recognition. Face recognition is a revolutionary technology and it can be used in a wide range of applications such as:

• People tagging in photos
• Gaming
• Price comparison
• Making mental notes
• Identifying TV shows
• Augmented reality
• Image searching
• Solving sudoku puzzles
• Security

Besides providing security through camera-based applications at both day and night, by displaying videos on a screen or recording video using a hard disk, face detection can be used to identify human faces among moving objects (people, animals, vehicles, and so on) within the field of view of the camera...

Raspberry Pi camera module

The Raspberry Pi camera module (Figure 7-1) helps you to make still photographs and high definition videos with all models of Raspberry Pi. The latest version of the camera module is V2.1 at the time of writing this book, but all previous versions should work with this project. The camera module offers the following features: Sony IMX219 8-megapixel sensor:

• Supports 1080p30, 720p60, and VGA90 video modes
• Supports still capture

Figure 7-1: Raspberry Pi camera module V2.1. Image credits: SparkFun electronics. https://www.flickr.com/photos/sparkfun/9367415640

Connecting the camera module to the Raspberry Pi

All Raspberry Pi models offer a CSI port to attach the camera module through a 15 cm ribbon cable. Figure 7-2 shows a camera module...

OpenCV

The Raspberry Pi camera module can be accessed through the MMAL, V4L APIs, and some third-party libraries, including the Picamera Python library. When you need to access the camera module with Java, a well-written, feature-rich library is available, which is OpenCV.

OpenCV is currently available for Windows, Linux/Mac, Android, and iOS as pre-built libraries. However, you can build the OpenCV library for your preferred development environment using the source code. For Raspberry Pi, we will be building the OpenCV library to work with Java in a few easy steps.

Downloading and installing OpenCV on Windows

In order to write Java applications with OpenCV, first you have to download and install Open CV on your computer. The following steps explain how to do this...

Working with video

as the key source for security purposes, rather than the still images. The video quality extends to full HD, which is 1080p, at a frame rate of 30.

It's time to work with video capture, which can be used as the key source for security purposes, rather than the still images. The video quality extends to full HD, which is 1080p, at a frame rate of 30. Listing 7-2 shows a Java program that can be used to capture video using the Raspberry Pi camera module and save a snapshot to the SD card. Replace Listing 7-1 with this:

Listing 7-2: SecurityCamera.java

package com.packt.B05688.chapter7; 

import org.opencv.core.Core; 
import org.opencv.core.Mat; 
import org.opencv.imgcodecs.Imgcodecs; 
import org.opencv.videoio.VideoCapture; 

public class SecurityCamera { 

 static { 
 System.loadLibrary(Core.NATIVE_LIBRARY_NAME); 
 } 

 public static void main(String[] args) { 
 Core.setErrorVerbosity(false); 
 VideoCapture...

Summary

In this chapter, you learned how to use the camera module with a Raspberry Pi to capture video as a series of frames, and process them to detect human faces using a cascade classifier, which is haarcascade_frontalface_alt.xml. You can improve this application by displaying the detected person's name, along with the highlighting rectangle, sending an email or SMS alert, or implementing a web stream to access the video through the network and display the video on a web page.