Week 8: Make-through

• Components:

Raspberry Pi 3

Raspberry bi compatible camera

bread board and some jumpers

our favorite tiny blue servo motor

We used open CV library as mentioned above and Haar cascade face detection classifier file (.xml).

Haar Cascade is a machine learning object detection algorithm proposed by Paul Viola and Michael Jones in their paper “Rapid Object Detection using a Boosted Cascade of Simple Features” in 2001

As you just read haarcascade classifiers are just too old and it's only used for educational purposes nowadays

Four libraries were used in this code:

PiCamera for using Raspberry pi camera

Picamera.Array for frame capturing

import time For sleep functions (similar to delay in arduino)

import cv2 for to use classifiers and do computer vision applicanions

In these lines we setup the frame dimensions, resolution, framrate and used classifiers.

In these lines we setup the 12th GPIO pin as an output pin to use it controlling the servo motor, then we define a PWM function to use it further in moving our servo

This the main loop that gives us our main functions, frame capturing+moving the servo.

first it changes the color format to grayscale which make it easier for the computer to analyse the image.

There the image is represented in a 2 dimensional array (x,y).

and as faces are detected, a rectangle is formed around it, as this center is moved a certain distance, servo1 function is called to move the motor