Skip to content

Latest commit

 

History

History

part-3

README.md

Part 3: Webcam tracking using Sensor Mezzanine

Track the known face infront of Webcam using Sensor Mezzanine and Dragonboard410C. Following are the steps necessary to acheive Webcam tracking.

If you would like to read more about this project, please visit this blog written by Manivannan Sadhasivam.

Table of Contents

1) Hardware

1.1 Hardware requirements

1.2 Hardware setup

  • DragonBoard 410 is powered off
  • Assemble Pan and Tilt servo by following the instructions here
  • Mount USB webcam on servo assembly and connect to Sensors Mezzanine according to the following connections:
    • Pin 9 —-> Servo controlling X axis
    • Pin 10 —-> Servo controlling Y axis
    • 5v (VIO) —-> Servo Vcc
    • Gnd —-> Servo Gnd Connections Setup
  • Connect Sensors Mezzanine
  • Connect USB webcam
  • Connect I/O devices (Monitor, Keyboard, etc...)
  • Power on your DragonBoard 410c with 96Boards compliant power supply

2) Software

2.1 Operating System

2.2 Package Dependencies

Installed the dependencies mentioned in Part-2 along with the following:

$ sudo easy_install pyserial

2.3 Arduino Programming

$ git clone https://github.com/96boards/projects.git
$ cd projects/home-surveillance/part-3

Open facetrack.ino using Arduino IDE and upload the code to Sensor Mezzanine with the following settings:

  • Board ----> Arduino Uno
  • Serial Port ----> /dev/tty96B0
  • Programmer ----> AVRISP mkII

3) Project Execution

Place the trained dataset and haarcascade_frontalface_default.xml created in Part-2 to current directory. Also, update the USB webcam video source as specified in Part-2

$ cp -r ../part-2/trainer ../part-2/haarcascade_frontalface_default.xml ./
$ sudo python facetrack.py

The above script will track the known face using webcam mounted on Servo Pan/Tilt system. The servo system will adjust it's position continuously to center the known face.

Script will prompt the user to enter the Name of the person that needs to be tracked. Type the name and press Enter.

Link to Blog