This really isn’t a tutorial on object tracking as the title promises. It’s a (good) tutorial for visual object detection and then a single-line black box function that does all the tracking with no explanation of its operation, just it’s use.
Dude! Not only have you helped me to finally set up opencv but you have even made my first steps easier in my own project with your demo. ❤ Hint to all who have troubles with the namespace in build.gradle: try to adjust the gradle plugin version under project structure > project. 7.2 worked for me.
Thanks for sharing the knowledge and resources. For the information of the fellow viewers/commenters.. I implemented his code in VS Code in Windows 10. 2 errors and following had to be done as it threw some error when I attempted to run it: 1. install "pip install tracker" --> Refresh VS code 2. Make sure that the specified folder path of the video and code are same. Files --> Open Folder--> Object-Tracking.
Hello, for those following the tutorial, a good enough roi can be ```python roi = frame[int(height/3):, int(width/2)-200:int(width/2)+200] ``` Good tutorial btw.
The way this person is delivering these complicated stuffs regarding Open CV and OT, hats off man for your contribution towards the TH-cam Community. Do try to make a playlist or video regarding implementation of GANNs tensorflow models along with Opened CV, like Real Time image to image translation using Cycle GANNs. By the way lots if love from this Subscriber 😊
Thank you so much men! I do appreciate your kindness and efforts for sharing this great script even for free. It helps a lot, especillay for many students who are with empty pockets! Gracias!
I am from Vietnam and I'm going to research to count the number of people in CCTV camera but I don't know how to track people. Then, I watch your video and I really like it. It's so helpful for me. I'm so grateful to you. Thank you so much
Hello . Great explanation, thank you very much. @00:10:44, It is clear that You filtered using area. But I wonder if it is possible to use some of "speed" threshold or a good combination (compound condition) instead of it???
Very interesting video. Is there some way to get the time that a vehicle pass through two lines in the highway and register this times in a excel sheet?
Sir where is tracker.py Is that pip install django-request-tracker If it is so then i am getting error saying Error in module and not in pip Please guide me
Awesome! I use the same method for tracking "alpha numeric " raised numbers and letters on the face of the front dashboard. My problem is that I have to process the images and then convert them for analysis. Do you have any cellular applications? Or desktop?
Can I change some line of code to only detect a specific moving object with a specific color; not all moving objects only the moving object with that specific color?
@@gsunil3697 Yeah, you can use the area of the rectangle as a paramter in any way, but it tends to be inneficient cuz you dont know what type of vehicles you can get in the image and the cam angle distorts the size
@@braiandeivid bro i have a code for vehicles detection but I want to classify them based on their size.but i am getting lot of errors .can you help me plz.....
Hi, I tried another video that has a stopped car and slow cars but this code can not track these cases...any help? from where I can start to fix this issue? thank you
Very nice video! But I found that the tracker keep assigning new ID to stationary objects or objects that is not moving fast enough. Do you have any tracker recommend that solve the problem? Many thanks!
Hi, the tracker.py file uses Eucledian Distance to track and generate new ids and it's only reliable for simple tracking tasks. For an accurate and reliable tracking you should integrate more advanced algorithms. On my course Object Detection (opencv and deep learning) I dedicated and entire module to object tarcking.
Hi i need your help have these punk neighbors who have twicked an algorithm object tracking following me around remotely using my voice and facial recognition. How do i stop them from following me...this is crazy i need helpnim a woman alone and they're abusing us. God bless and thank you
Is there a way to optimize tracker class? In my case, I will use it on a robot which will operate on the roads. Camera will vibrate so I want it to keep vehicles' ids even though the camera vibrates.
A very nice tutorial, Thanks. I would like to save images for every single car (instead of every frame) inside the rectangle from the video. How to do that please?
I would assume you have to set a way to save the video from once it detects a specific car or motorcycle and end said video clip recording once that object is gone.
Nope, this is a very simple object tracking to be considered only good as an exercise. To make things reliable and precise on CCTV it's necessary to use object detection with deep learning and some algorithms for a reliable tracking when there is occlusion
Hi... Thank you for the tutorial.. It's impressive. I have one question. Can we make the animation of moving objects a function of time? the program will run and continuously animation will run??? Thank you
Hey Soumitra, I'm actually trying to do that too and I wondered if you came to a solution. :) I know I'm probably to late asking this, but I will be very happy if you reply. Have a great day
Sir you said we should check one of your program "Object detection with opencv and Deep learning " but i can't find the lecture.. Please can you send me the TH-cam link please 🙏🙏🙏
Hello, it is possible to track people? if yes, how we can detect if someone is with frontal position or back position do determinate direction of the movement? this is because i need to count if leave some place or enter to the place
Yes it's possible, this is something I do on my Computer Vision course, and something I help companies with. You will need to use deep learning to detect the people + object tracking algorithms to keep track of them
really cool tutorial, thank you! One question tho; in the video example the detection number 12 gets updated to 13, although its the same object - how do you deal with these cases? is there a good way to clean up these detections so you don't tag the same object multiple times and increment its number when its not necessary? thanks!
this should be considered only as an exercise to learn more about tracking and computer vision. For a more robust tracking, it's necessary to integrate deep learning object detection and tracking algorithms.
Sir, thank you for the video with a detailed explanation. If I want the program to execute on two different videos at a time is that possible. If yes please tell the way
thank you very much, great Tutorial! My Problem is that the Object has many Shadows, reflected from the Street and from the Window from the Car. do you have there a solution for?
Yes,. Consider this is just one of the many approches possible. The most reliable one will be integrating deep learning with Opencv and object tracking, and you can have an accuracy close to 100%
🔥Learn how to build your own AI vision solutions: pysource.com/community
Today I installed opencv, it was a long day of tutorials but this one was super super clear!
Thanks a lot!
super charge
This really isn’t a tutorial on object tracking as the title promises. It’s a (good) tutorial for visual object detection and then a single-line black box function that does all the tracking with no explanation of its operation, just it’s use.
Can Using visual video code to run?
Dude! Not only have you helped me to finally set up opencv but you have even made my first steps easier in my own project with your demo. ❤
Hint to all who have troubles with the namespace in build.gradle: try to adjust the gradle plugin version under project structure > project. 7.2 worked for me.
Thanks for sharing the knowledge and resources.
For the information of the fellow viewers/commenters..
I implemented his code in VS Code in Windows 10. 2 errors and following had to be done as it threw some error when I attempted to run it:
1. install "pip install tracker" --> Refresh VS code
2. Make sure that the specified folder path of the video and code are same. Files --> Open Folder--> Object-Tracking.
Hello, for those following the tutorial, a good enough roi can be
```python
roi = frame[int(height/3):, int(width/2)-200:int(width/2)+200]
```
Good tutorial btw.
Thanks, to you Sergio Canu,
you have resolved much of our issues.
thanks to you from the core of our hearts.
The way this person is delivering these complicated stuffs regarding Open CV and OT, hats off man for your contribution towards the TH-cam Community. Do try to make a playlist or video regarding implementation of GANNs tensorflow models along with Opened CV, like Real Time image to image translation using Cycle GANNs.
By the way lots if love from this Subscriber 😊
i'm newbie to opencv , u explanations are far better than my lectures....thansks a lot sir.....
this was really awesome and thank you for the code
Thank you so much men! I do appreciate your kindness and efforts for sharing this great script even for free. It helps a lot, especillay for many students who are with empty pockets! Gracias!
Object Tracking starts at 20:25
Hello !! From where can I get tracker.py
BEST youtube channel!
This was my first AI program in python. Thanks a lot for teaching this. 🙏🙏🙏
Best tutorial ever!!!! Thanks a lot for your time and clear explanation, cheers!!!
Thank you for explaining each and every line of code clearly. Great help. Keep uploading videos like this
I am from Vietnam and I'm going to research to count the number of people in CCTV camera but I don't know how to track people. Then, I watch your video and I really like it. It's so helpful for me.
I'm so grateful to you. Thank you so much
Thanks man! This helps me to complete my assignment on object tracking using OpenCV
It’s great, it makes me think python is awesome, and I start to like it.Thankyou！
Wow this was quite useful and well explained. Thank you and God bless your work!!!
Bravo Sergio! Maestro!
Thank you for the video sir, I have spent weeks to do this traffic counting and none that I found work for me. Thank you so much for the tutorial!
perfect and much needed for me to understand OPENCV
From where I can get tracker.py file
We used it for our study. Thank you.
我來自台灣，謝謝你的教導，感謝!!
Hello . Great explanation, thank you very much.
@00:10:44, It is clear that You filtered using area.
But I wonder if it is possible to use some of "speed" threshold or a good combination (compound condition) instead of it???
Thank you very much for making this video; concepts are clearly explained.
Awesome video, where can i get traffic footage like that
Very interesting video. Is there some way to get the time that a vehicle pass through two lines in the highway and register this times in a excel sheet?
Thanks I finally got a good starter tutorial
Thanks, Are you able to count the demonstrators of a march ?
Sir where is tracker.py
Is that pip install django-request-tracker
If it is so then i am getting error saying
Error in module and not in pip
Please guide me
It is a custom made function
thanks for great video, does it work for moving camera and stable object too?
Awesome! I use the same method for tracking "alpha numeric " raised numbers and letters on the face of the front dashboard. My problem is that I have to process the images and then convert them for analysis. Do you have any cellular applications? Or desktop?
Thanks for sharing! Keep up the good work 😄
;)
Can I change some line of code to only detect a specific moving object with a specific color; not all moving objects only the moving object with that specific color?
Yes, you can do everything
Is it possible to classify vehicles based on their size
@@gsunil3697 Yeah, you can use the area of the rectangle as a paramter in any way, but it tends to be inneficient cuz you dont know what type of vehicles you can get in the image and the cam angle distorts the size
@@braiandeivid bro i have a code for vehicles detection but I want to classify them based on their size.but i am getting lot of errors .can you help me plz.....
is there some code or resource to leave a 'tail' that traces the path of the rider?
Where is the tracker file
It's not in the link below
is this a different method that we approach object tracking by using different functions in opencv, do you find it yourself?
its great ... free source codes. keep posting free codes.
Thank🎉 from Thailand.
Hi, I tried another video that has a stopped car and slow cars but this code can not track these cases...any help? from where I can start to fix this issue? thank you
Awesome ... Great Work
Are x and y coordinates mean the center point of the rectangle, or are they the corner of the rectangle?
I want to know the exact date and time, when a vehicle entered the frame of video? How do I do it? Supposing we are feeding live video from a camera?
Hello, is it possible to create software that identifies the speed of an object in a video? Please
Subscribed! A very beautiful explanation and examples
Hello can you help me to find camera coordinate and contour position on a image or realtime video
Is it possible to use an external camera as a video source?
What if I wanna do this live. Like from a streaming camera ? Or from live video playback
Thanks for the insight, you have taken my python-programming to another higher level!! Great fun too.
How can I get tracker module for euclideandistracker ? I cannot find it
Thank you for the tutoral, very good. i only want to ask one question. How can i restart to zero the id counter?
How do you connect your program with a camera
Hey!
Thanks a lot for the video! Very helpful!
Tell me, how can I record the coordinates of landmarks in a json file?
can i use vlc window panel as a motion capture ?. cuz i want be able to change the video source :)
Thank you so much men! I want to study follow your ! 👍
Thank you. Super lecture. Very insteresting.
Very nice video! But I found that the tracker keep assigning new ID to stationary objects or objects that is not moving fast enough. Do you have any tracker recommend that solve the problem? Many thanks!
Hi, the tracker.py file uses Eucledian Distance to track and generate new ids and it's only reliable for simple tracking tasks.
For an accurate and reliable tracking you should integrate more advanced algorithms.
On my course Object Detection (opencv and deep learning) I dedicated and entire module to object tarcking.
your tutorial method is wonderful ...
Can you do a lesson on calculating the size of objects such as fruits and vegetables . Thanks
Solid explanation to initiate into the field
Hi i need your help have these punk neighbors who have twicked an algorithm object tracking following me around remotely using my voice and facial recognition. How do i stop them from following me...this is crazy i need helpnim a woman alone and they're abusing us. God bless and thank you
Is there a way to optimize tracker class? In my case, I will use it on a robot which will operate on the roads. Camera will vibrate so I want it to keep vehicles' ids even though the camera vibrates.
A very nice tutorial, Thanks. I would like to save images for every single car (instead of every frame) inside the rectangle from the video. How to do that please?
I would assume you have to set a way to save the video from once it detects a specific car or motorcycle and end said video clip recording once that object is gone.
I really enjoy watching your videos and they are quite helpful at same time. thanks for sharing this with us
hi. great tutorial.
Can we apply this technique to our cctvs?
Thanks.
Nope, this is a very simple object tracking to be considered only good as an exercise. To make things reliable and precise on CCTV it's necessary to use object detection with deep learning and some algorithms for a reliable tracking when there is occlusion
Thank you for your detailed explanation on object Tracking :)
Sir do you have any project for hand gesture recognition ? Thanks
Nice video
How do I implement this using YOLO v5
how can i store the output in an excel file?
thnx!
Can you explain more about the working of tracker.py file. Thanks u
Where is tracker.py file
Can you please tell how to add line that u speak off in last
Hi...
Thank you for the tutorial.. It's impressive. I have one question.
Can we make the animation of moving objects a function of time? the program will run and continuously animation will run???
Thank you
Hey Soumitra,
I'm actually trying to do that too and I wondered if you came to a solution. :)
I know I'm probably to late asking this, but I will be very happy if you reply.
Have a great day
Sir you said we should check one of your program "Object detection with opencv and Deep learning " but i can't find the lecture.. Please can you send me the TH-cam link please 🙏🙏🙏
Hello, it is possible to track people? if yes, how we can detect if someone is with frontal position or back position do determinate direction of the movement? this is because i need to count if leave some place or enter to the place
Yes it's possible, this is something I do on my Computer Vision course, and something I help companies with.
You will need to use deep learning to detect the people + object tracking algorithms to keep track of them
Please I can't find the link to download those files required for this lesson, can anyone help me?
pls how can i get the velocity of the moving objects
where to download the video?
Anyone having trouble with receiving the email with the necessary files on Pysource website?
May I ask how to export the location of the object (x,y,w,h) into a CSV file?
This is just great. Many many thanks 👍
very very awesome, thank you
How can I detect videos that have been downloaded?
every time I run it only comes out of the camera view, it doesn't immediately detect the video
can we do this in google cloud
Bro,Is it possible to classify vehicles based on their size.plz help me.
Can you do this for mobile camera?
sir how to get x,y with respect to time foe a single object that should be given as input
Dude, thanks for this video!
really cool tutorial, thank you!
One question tho; in the video example the detection number 12 gets updated to 13, although its the same object - how do you deal with these cases?
is there a good way to clean up these detections so you don't tag the same object multiple times and increment its number when its not necessary?
thanks!
this should be considered only as an exercise to learn more about tracking and computer vision. For a more robust tracking, it's necessary to integrate deep learning object detection and tracking algorithms.
@@pysource-com gotcha :D thanks for the reply :)
Congratulations!
Can you do the same for the number of people demonstrating? Thank you
how can i track the bot using moving camera that is attached into the same bot? any idea??
Thank you so much, please make more tutorials like tNice tutorials. You explain better than any other guide I've seen on youtube, very clear and
hello sir do you have a tutorial for detecting not moving object
and also Counting it
wow, you explain very well, trank u bro!
Sir, thank you for the video with a detailed explanation.
If I want the program to execute on two different videos at a time is that possible.
If yes please tell the way
Thanks a lot for this amazing tutorial !!!
thank you very much, great Tutorial! My Problem is that the Object has many Shadows, reflected from the Street and from the Window from the Car. do you have there a solution for?
Yes,. Consider this is just one of the many approches possible.
The most reliable one will be integrating deep learning with Opencv and object tracking, and you can have an accuracy close to 100%
@@pysource-com so this example is not based on deep learning, right?