Object Detection using OpenCV Python in 15 Minutes! Coding Tutorial #python #beginners

Summary

This tutorial introduces viewers to object detection using OpenCV and Python. The presenter demonstrates how to install necessary libraries, access a camera feed, and identify various objects in real-time. The video also covers how to use the gtts and play sound libraries to make the computer vocalize the detected objects, creating an interactive and informative experience for the audience.

Takeaways

• 😀 The tutorial focuses on object detection using OpenCV, a popular computer vision library.
• 🛠️ The presenter guides through the installation of necessary libraries, including `opencv-contrib-python`, `cvlib`, `gtts`, and `play-sound`.
• 🔎 `opencv-contrib-python` is preferred over `opencv-python` for its additional libraries that enhance functionality.
• 📱 The script demonstrates real-time object detection using the computer's webcam, identifying various objects like an apple, orange, and cell phone.
• 🎯 The `cvlib` library is utilized for its pre-trained models to recognize common objects within the video frames.
• 🗣️ The tutorial includes a feature to convert detected objects into spoken words using Google Text-to-Speech (gtts).
• 🔊 The `play-sound` library is integrated to efficiently play the synthesized speech.
• 📝 The script maintains a list of unique detected objects to avoid repetition in the output.
• 📑 The tutorial concludes with a function that converts the list of detected objects into a natural-sounding sentence and plays it aloud.
• 🎉 The presenter encourages user interaction through comments, likes, and subscriptions for further tutorials.

Q & A

What is the main focus of the tutorial?

The main focus of the tutorial is to demonstrate how to use OpenCV for object detection, allowing the computer to identify and announce different objects seen through a camera feed.

Why is OpenCV-contrib-python used instead of OpenCV-python?

OpenCV-contrib-python is used because it contains additional libraries beyond the basic modules of OpenCV-python, providing more functionality for advanced tasks such as object detection.

What libraries are installed for object detection in the tutorial?

The tutorial installs 'opencv-contrib-python', 'cvlib', 'gtts', and 'play-sound' libraries to handle object detection, text-to-speech conversion, and audio playback, respectively.

How does the tutorial handle real-time object detection?

The tutorial accesses the camera using 'cv2.VideoCapture' and processes each frame in a loop to detect objects in real-time, then draws boxes and labels around the detected objects.

What function is used to draw boxes around detected objects?

The 'draw_box' function from 'cvlib.object_detection' is used to draw boxes around the detected objects in the video frames.

How is the list of detected objects managed to avoid duplicates?

The tutorial uses a for loop to check if an item is already in the 'labels' list before appending it, ensuring each object is only announced once.

What is the purpose of the 'speech' function defined in the tutorial?

The 'speech' function is used to convert the list of detected objects into a natural-sounding sentence and then use the 'gtts' library to convert the text into speech.

How does the tutorial ensure a more natural pause in the spoken output?

The tutorial uses string interpolation to create a sentence with 'and' and commas, which are then joined into a single string to ensure natural pauses when the computer speaks the detected objects.

What is the significance of creating a 'sounds' directory in the project?

The 'sounds' directory is created to store the audio files generated by the 'gtts' library, which are then played using the 'play-sound' library.

How does the tutorial handle user interaction to stop the object detection process?

The tutorial uses a 'cv2.waitKey' function to check if the user presses the 'q' key, which, if pressed, breaks the loop and stops the object detection process.

Outlines

00:00

😀 Introduction to Object Detection with OpenCV

The video begins with an introduction to object detection using OpenCV. The presenter demonstrates holding up various objects such as an apple, an orange, and a cell phone, and explains the goal of getting the computer to verbally identify these objects within a video frame. The presenter then proceeds to explain the first step, which is to install the necessary dependencies. They guide the viewers through the installation of 'opencv-contrib-python' and 'cvlib' using pip, highlighting the additional libraries provided by 'opencv-contrib-python' over the standard 'opencv-python'. The presenter also addresses potential warnings about updating pip and suggests resolving them promptly.

05:01

🛠️ Setting Up Object Detection and Speech Synthesis

In this section, the presenter continues the setup process by installing additional libraries for speech synthesis and efficient sound playback. They use 'gtts' for text-to-speech conversion and 'play-sound' for playing the synthesized audio. The presenter then imports necessary modules from OpenCV, cvlib, and the other installed libraries. They explain the process of accessing the computer's camera to capture live video for real-time object detection. The script includes a loop to continuously read frames from the video capture and use cvlib to detect common objects, drawing boxes and labels around the detected objects.

10:02

🔍 Detecting and Storing Object Labels

The presenter explains how to process the detected object labels to avoid duplicates in the list. They create a list called 'labels' and use a for loop to append unique labels to this list. The script checks if an item is already in the 'labels' list before adding it, ensuring each object is only listed once. The presenter then demonstrates testing the list by printing its contents, which shows the successful detection of unique objects like a person and a tie.

15:04

🗣️ Implementing Speech Output for Detected Objects

The final part of the video focuses on converting the list of detected objects into spoken words. The presenter creates a function called 'speech' that takes text input, converts it to speech using gtts, saves the audio to a file, and plays it using the 'play-sound' library. They demonstrate how to construct a natural-sounding sentence from the list of labels and then use the 'speech' function to vocalize the findings. The video concludes with a live demonstration of the complete setup, where the computer successfully identifies objects in the video feed and verbally announces them, followed by a call to action for viewers to engage with the content and a wrap-up of the tutorial.

Keywords

💡OpenCV

OpenCV, which stands for Open Source Computer Vision Library, is an open-source computer vision and machine learning software library. It is used for various image and video analysis tasks, including object detection, as demonstrated in the video. The script mentions installing OpenCV with additional libraries to enhance its capabilities for object detection.

💡Object Detection

Object detection is a computer vision technique that involves identifying and locating multiple objects in an image or video. In the video, the presenter uses OpenCV to detect various objects like apples, oranges, and cell phones, and then uses text-to-speech to announce what the computer 'sees'.

💡cv2.VideoCapture

cv2.VideoCapture is a function in OpenCV used to capture video from a camera. In the script, it is used to access the computer's camera to provide a live feed for real-time object detection.

💡cv2.imread

cv2.imread is a function in OpenCV used to read images from a file. Although the script mentions initially using this function with a specific image, the focus shifts to using live video feed for object detection.

💡cv2.imshow

cv2.imshow is a function in OpenCV that is used to display an image in a window. The script uses this function to show the live video feed with detected objects and drawn boxes around them.

💡cv2.waitKey

cv2.waitKey is a function in OpenCV that waits for a key event from the user. In the script, it is used to allow the user to exit the video feed by pressing a specified key, such as 'q'.

💡gtts

gtts stands for Google Text-to-Speech, a Python library that converts text to speech. The script uses gtts to convert the detected object labels into spoken words, which are then played back to the user.

💡play sound

play sound is a Python library used to play audio files. In the context of the video, it is used in conjunction with gtts to play the audio output of the text-to-speech conversion.

💡cvlib

cvlib is a library for computer vision tasks, including object detection. The script mentions installing cvlib to use its pre-trained models for detecting common objects in the video feed.

💡String Interpolation

String interpolation is a method of embedding expressions within a string, which are evaluated and then replaced with their values. In the script, it is used to dynamically create sentences that describe the detected objects, which are then spoken by the computer.

💡List Comprehension

List comprehension is a concise way to create lists in Python. The script uses list comprehension to create a list of unique detected objects, ensuring that each object is only announced once.

Highlights

Introduction to object detection using OpenCV.

Installation of dependencies including OpenCV and cv-lib.

Explanation of the difference between OpenCV Python and OpenCV-contrib-python.

Installation of gtts and playsound libraries for text-to-speech functionality.

Importing necessary libraries for object detection and speech.

Accessing the camera for a live feed of object detection.

Loop to process each frame from the video capture for object detection.

Using cv-lib to detect common objects and draw boxes around them.

Displaying the detected objects with labels in real-time.

Creating a list to store unique labels of detected objects.

Explanation of avoiding duplicate entries in the labels list.

Printing the list of detected objects to verify the functionality.

Creating a more natural-sounding sentence from the list of detected objects.

Using string interpolation to format the detected objects into a coherent sentence.

Defining a function to convert text to speech using gtts and playsound.

Saving the generated speech as an MP3 file.

Playing the generated speech to confirm the object detection results.

Final demonstration of the object detection and speech synthesis working together.

Encouragement for viewers to ask questions and subscribe for more tutorials.

Transcripts

[Music]

i found a person and a tayachara apple a

doughnut orange a cell phone hey guys

welcome to another tutorial with opencv

today we are going to look into object

detection so i can just hold up an apple

or an orange

heck even my cell phone lots of

different objects and i'm going to get

my computer to tell me out loud using

its voice what it saw inside of the

frame that's enough introduction let's

get right into

[Music]

it so the first thing that we need to do

is install our dependencies so we are

going to import a couple of libraries

the first thing that we're going to

import is opencv contribute python so

we're going to say

pip

install open cv dash contrib dash python

and some people ask me why i use this

lately instead of just open cv python

it's because opencv python does contain

the main components that you need for

the basic modules of opencv but with

opencv can trip python it's going to

contain some extra libraries so that we

have a little bit extra to work with and

so we're just going to hit enter and

that's going to start installing okay

and if you ever get a warning like i did

where you need to update pip or anything

you can go ahead and do that as well i'm

just going to copy and paste that that

should be just really quick

and there we go perfect after we have

that installed we are now going to go

ahead and install c-lib

so pip install

cv lib okay and we're going to be using

this for our object detection so there's

a library that's already learned what

certain objects are so we're just going

to install that

and depending on your internet

connection that should be rather quick

very good and then we're going to also

allow our computer to

say out loud what it saw so if it sees

me it'll say i saw a person or i saw an

apple an orange so on and so forth so we

we're going to import just a couple more

things we're going to say pip install

gtts

space

play sound whoops

play

sound like that and then finally we are

going to

install

pi object c which is going to help with

that sound be a little bit more

efficient so i'm going to say

pip

3

install

capital p y capital o b j capital c

okay so that will allow play sound to be

a little bit more efficient so i already

have a couple of those installed

installs just gonna say already

satisfied for you it will probably stay

successful if you have any errors just

go back rewind make sure that you typed

everything correctly

otherwise let's move on

i'm just gonna slide down my window here

and i'm going to now import

cv2

import

cv lib as cv

and then from

cvliv.object

detection

import

draw box so it's going to be drawing a

box around our objects for us so make

sure you have two b's for box b b o x

and then we're gonna say from g t t s

import

g capital t t s oops i said g t a let me

do g t t s then finally from play

sound

import

play

sound so there's one two three four five

lines of imports but that is everything

that we're going to be using for this

video so if you need a little bit more

time you can go ahead and pause the

video and continue that

three days later so what we want to do

now is now access our camera now

originally when i was first building

this and testing things out i was just

having it bring in a specific image i'm

just going to look at objects in the

image but i wanted this to be live so it

has a live feed and we can detect all

the objects in the live feed instead so

we're going to access our cameras

so i'm going to say video equals

cv2 dot video

capture

and that takes an index now for most of

you it might be index zero but my web

camera that i want to be using instead

which is a lot more higher quality is

that index one

so you can just mess with those indexes

as you please

but we're going to start with that and

now we're going to say while

true

i'm now going to use my video capture

and i'm going to unpack

each frame into a

variable called frame so what we're

going to do is ret

comma

frame

equals our video

dot read

so unpack that so now we're going

through each frame

and now we're going to use that bb box

where it's going to be seeing the

objects it's going to draw a box around

it and we're also going to give it a

label next to the box to tell us what

the object is

so we're going to say bb box

comma label

and then

conf okay and comp is really just

identifying what the object is it's just

going to be returning some decimal

numbers really so i'm going to say cv

dot detect

common objects

and now we have to tell it where to get

those objects from

so we need to say get it from the frame

so that's going to be each frame from my

video feed and finally we're going to

draw that box so we're going to say

output

image

equals draw

box

and now we need to give dropbox the

frame

i want it to get the

box that it's going to be drawing around

and we're also going to put the label in

there and we'll stick conf in there very

good so now that we have that let's go

ahead and show the user what the image

looks like so i'm going to say cv2 dot i

am show

and we want to show them

the name of the window so i'll just call

this uh object

detection you can call that whatever you

want comma and now we've got to tell it

output image okay before we hit run

we're going to um

give this a wait key so i'm going to say

if

cv2 dot weight key

delay of one and

we're going to check to see if the user

is clicking a certain button

you can say whatever button you want but

i'm going to say if the user

clicks on

q some people like the space bar you can

just do a space

do whatever you want but i'm going to

say if the user says hits q i want you

to break out of this loop

and after i hit q it breaks out of that

window so very good so as you could see

that was already detecting me as a

person even detected this as a tie

that's already working so what i want

this to do now is i want my program to

take each of those labels that it finds

in my screen

and i want it to append or add to a list

so that i have that list of data so what

we're going to do now is we're going to

make a list called labels so let's come

up here and we'll call this labels

make sure this is outside of your loop

so it doesn't accidentally rename itself

inside the loop and just erase all the

data

um and what we're going to say is

we'll do a for loop we're going to say

for

item in label

if item in labels

then we're just going to have it pass

so that means if if it already found a

tie it's going to be checking multiple

images it's going to be checking for

objects in each frame

and so it's going to

say maybe like a thousand ties i don't

want it to do that i'm just going to say

if you find a tie in there then go ahead

and put it in the list but if tai is

already in the list don't add it to the

list so it's only going to say tie one

time and you can alter this if you'd

like

but this is the way i'm going to do it

um

if it's not already in the list then i

want you to

labels.append i want you to append

that item

so that item will be added to this list

called labels

and just to test that out

i'm going to

come down here and print

labels

and let's see if that works

okay and as you can see here is my list

called labels that i just printed so i

found a person and it found a tie so

very good uh i know that this is working

because if i didn't it would be the same

person

a thousand times in the tie a thousand

times but it's only gonna do it once

because of this code here so very good

so what i want this to do now is i

wanted to take this

data called labels so what i'm going to

do now is write code

using string interpolation to tell me

what it found

more logically for example i wanted to

say something like i found a orange a

person a book

a tie a cell phone an apple so on and so

forth

and so

how i'm going to do that is i'm going to

create a for loop for label

in labels

i want this to check to see if this is

the first time it's reading out loud a

label so i'm going to create some kind

of iterator so i'm going to say i equals

0 and i'm going to say here

if i

is 0

then i want this to actually append to a

list because i want this to

sound a little bit more natural when it

says it out loud so i'm going to say new

sentence

equals an empty list so i'm going to say

if i zero

then new

sentence dot append

that means to add and i want it to add

i'll use some string interpolation here

i found a

i'll put label there so if it's found a

person will say i found a person

and i'll do a comma and comma

that'll give this speech out loud to

give a little bit of a pause

so we'll see how that goes

and

then i'm gonna say if it is

not equal to zero

then new sentence dot append and i'll

use string interpolation again

and i'm gonna say

uh

label like so

and then once that is done we are going

to increment

i so i'm going to say

i

plus equals one so the first thing it

finds

i is it going to be zero so the first

thing it finds is going to say i found a

hat

and

a person

a book an apple so on and so forth so

very good with that uh and just to make

sure that this is all going to be all in

one string for our speech

to work properly after this let's go

ahead and use the join function so i'm

just going to say print

space dot join

new

sentence so that will turn this list

into a string so let's go ahead and test

that out

perfect so after hitting run and after

it found those things check it out it

added each of those things to the list

so said i found a person and a tie a

orange a chair a donut a apple again

this might look kind of weird with the

commas but it's going to help with the

pauses when the computer says it out

loud so very good if we have that

working just fine then let's go ahead

and add our speech part of this so

towards the top of our project

i'm going to now add

our speech

so let's go ahead and

define a function so def

we'll call this speech

and it's going to receive some text and

if you've seen my

virtual assistant where you build your

own siri or alexa this is the same exact

function that we're going to be using

so we're going to say

print

that text because we want to be able to

see it as well and i also want to set my

language to whatever i want i'm going to

set mine to

i will call yeah we'll say english so e

n

if you want to do spanish it's e es or

japanese it's ja you can always look

those up on google if you'd like to but

we'll do english for now

and then we'll give this some output so

output

equals now we're going to use gtts so

gtts which takes some text

which is going to be equal to the text

that we're going to send it

comma

now it's looking for the property of

language and so that will be our

language

and then finally it's going to ask how

fast we want to go so i'm going to say

slow equals false

just like that now what we need to do is

save

the output

into a file so what we need to do is we

need to save that audio

somewhere in our project so come into

your project and let's create a new

directory

and we're just going to call it

sounds so here's sounds i forgot to

change the name of the project so don't

worry about that but sounds is just

underneath that directory so

with that in place we can now save so

we're going to say output dot save

and now we're going to tell where to

save

so i'm going to say dot slash

dot slash sounds

and now call your file whatever you want

i'm just going to say output.mp3

says this will be a mp3 file and then

finally we're going to have it play the

sound

so we're going to use the play sound

library that we imported up here and

we're going to say

play sound which is gonna be that same

exact location

output dot mp3 and now we gotta send

whatever text we want over here

so down here we have a print instead of

print i'm going to say speech because

that's the name of our function

so that will send our string to our

function and it's going to take our text

and make it into actual speech it is

then saving

and then we're going to play that sound

let's see if this works

[Music]

[Music]

i found a person and a tayachara apple a

doughnut orange a cell phone

so hopefully you could have heard that

but it did say i found a person and a

tie an apple a orange a cell phone so

that is working great congratulations if

you just accomplished that that was

really cool pretty simple and if you

have any questions please let me know

down in the comments if you have any

requests please let me know don't forget

to drop a like and subscribe so that you

are notified of my next tutorial thank

you so much and happy coding

[Music]

you know