Duration 17:59

Traffic Signs Classification Using Convolution Neural Networks CNN | OPENCV Python

by Murtaza's Workshop - Robotics and AI

147 700 watched

0

2.2 K

Published 2 Jan 2020

Train and classify Traffic Signs using Convolutional neural networks This will be done using OpenCV in real-time using a simple webcam. CNN's have been gaining popularity in the past couple of years due to their ability to generalize and classify the data with high accuracy. In this video, we will train traffic signs with over 35000 images of 43 different classes with the help of TensorFlow and Keras. By the end of the video, I will share information that will help you classify your own data set. Info such as how long does it take to train and how much data of each class is required to have a good classification model. Code & Files: https://www.computervision.zone/courses/traffic-sign-classification/ Premium Courses: ✔️ Computer Vision Game Development Course: https://bit.ly/3ttLZ2s ✔️ Computer Vision with Arduino Course: https://bit.ly/3wzLB4m ✔️ Advanced Drone Programming Course: https://bit.ly/3qs3v5g ✔️ Learn to Build Computer Vision Mobile Apps: https://bit.ly/3uioY1J ✔️ Jetson Nano Premium Course: https://bit.ly/3L8uIlF Follow Me: TikTok: https://bit.ly/3Vo76OQ Facebook Group: https://bit.ly/3irDcb7 Discord: https://bit.ly/3JvyxAM Facebook Page: https://bit.ly/3IvpU7W Instagram : https://bit.ly/3NdGME3 Website: https://bit.ly/3ICFTS0 Github: https://bit.ly/3woU6PS Product Links: Recommend Webcam for Computer Vision: https://amzn.to/2MNtVKZ Budget Webcam: https://amzn.to/2ZP47Ug Computer Vision Robot Arm : https://amzn.to/3L1YacX Cheap Drone for OpenCV: https://amzn.to/2TZpsJy DC Motors + Wheels + Chassis: https://amzn.to/2SCZon3 DC Motors + Wheels: https://amzn.to/2QeEusw Arduino UNO: https://amzn.to/3Jwpz6h Motor Driver: https://amzn.to/35grl6x Battery: https://amzn.to/2Fadc0c Raspberry Pi 4 Best Starter Kit: https://amzn.to/3JvnEz1 Raspberry Pi Recommended Battery: https://amzn.to/2C0I9pl My Setup: Mouse: https://amzn.to/3tsx3BR Mechanical Keyboard: https://amzn.to/3JyVV0q Normal Keyboard: https://amzn.to/3L325WJ GPU: https://amzn.to/3NdzmjW CPU: https://amzn.to/3wsmhgI SSD: https://amzn.to/3wzY7AS MIC: https://amzn.to/3D43TMk Camera: https://amzn.to/36yvl90 3D Printer: https://amzn.to/3ipWNZ4 Sim Race: https://amzn.to/3IqfvKJ #ComputerVision #OpenCV #CVZone

Category

Show more

Tags

Artificial intelligence

Convolutional neural networks

Convolutional neural network

Artificial neural network

Convolutional neural network python

Neural networks

Cnn keras

Convolutional neural net keras

Deep learning keras

Keras tutorials

Traffic sign classification

Traffic sign opencv

Traffic sign keras

Traffic sign tensorflow

Tensorflow tutorial

Opencv tensor flow

Opencv keras pyhton

Traffic Signs Classification Using Convolution Neural Networks CNN

Comments - 674

@

@cyclone465O4 years ago Good job, very detailed stuff. Keep it up! 2
@

@danieldollive46602 years ago The gratitude for you work to help teach and really get us proficient in programing and process flow, You may change and enhance many of our lives.
@

@hackercop3 years ago Sir I find it much easier to learn when you code the program through the video and we follow along rather than you just explaning through it. Great video anyway, I'm learning a lot of the tricks. 8
@

@saramadani48872 years ago Awesome! Easy to understand. One request, I get the folder of ( myData, labels and source code of cv2) but can not find the source code of python that you introduced at the first. & I Follow the link in the description and enroll to get access to all code and files. ... 5
@

@merveozdas11932 years ago You değerce more subscriber, this is very inspiring channel 😍
@

@manochinnam4 years ago I like your videos and subscribed your channel. Keep post more videos Thanks !
@

@mohammadaideiramli9574 years ago Thanks Sir, I love your tutorial so much 1
@

@mustafaozcam4 years ago First of all, I respect you for these educational videos. How many problems will there be, I would be happy if you answer:
1) Can I do this training with my own different data?
2) I tried to do but I encountered the following error: "AssertionError: The dimesions of the Training images are wrong
"What is the solution to the problem?
3) Can you share the pictures of these videos?
Best regards. ... 2
@

@tanujak65773 years ago Thanks a lot for this video Murtaza, the code worked great! 1
@

@hasnainkhan29594 years ago Hello Sir, really loved the video thanks for this awesome video. Sir, I was thinking to re-implement it with SVM so can you guide me on how can I do this? Do I need to change the only classifier or other changes will be required also? I will be grateful. ... 2
@

@muhammadrezahaghiri3 years ago It is great, and I like how your computer fans went mad at 😂
@

@J11LM4 years ago Another question, would it be too difficult to implement the detection part to this project? For it to work in real life 5
@

@gufransabri90123 years ago I have never used keras. Where can i learn keras to understand this application better? is the deep learning fundamentals course on the DeepLizard channel good enough? 1
@

@Omundodoarthur4 years ago Amazing video! My congratulations and thanks for sharing. Do yiu have any video about object detection?
@

@unknown_34153 years ago Very good explanation has been done, I was trying to do this traffic sign project, but the line code (history=...) is not visible valudaty_data. Can you please tell me the complete line of it? It would b very helpful ... 2
@

@Hitesh-oz1dx4 years ago Hi, I was wondering where I could download the test data for the neural network as the link to your website if just showing an error 1
@

@vanquyet39859 months ago A successful project that I have accomplished and achieved excellent results
@

@michaelschmitt69314 years ago Thank you for this interesting video. How does it work under real conditions when you are driving in a car? Backlight? Darkness, only the lights of your car? Multiple signs at once?
@

@pggp92424 years ago Loving the tutorial so far. But I do have a question. For some reason tensorflow is defaulting to using my CPU. Anyway to force it to use my GPU? 1
@

@chiragjariwala32644 years ago I have one question!! There is too less image for train in class i copy all that picture and paste it again in that folder and make it around 1200 as other folders have .can train my model good? I mean copying picture will work? ...
@

@mustafa97radhwan3 years ago please : could you explain the right method to install tensorflow with pycharm
because there are many errors 2
@

@lolraiser4 years ago Hello, love the video <3, but i have a question , this will work if there are other things in the picture ? like if we have a picture of a streat and there is a stop sign, it will recognize it ? 2
@

@ClayAnimationDudelast month i was wondering if we could incorporate this with the same program used to run the robocar? 1
@

@longnguyenthanh75153 years ago Hi, i have watched your video, it's really excellent. How can i find the data ? Tks
@

@aakashsinghrawat33134 years ago Very well, got a question, cv2.resize(32,32) will resize the frame very little size.any alternative to get big frame?
@

@Manishkumar-ww4gm2 years ago I am working on making my own datasets for Indian traffic sign classifications.Can you suggest me any good idea.
@

@sontruongngoc56254 years ago Thank you for wonderful video. Can you tell me how I can create the dataSet? 1
@

@oumaimaelmellass82514 years ago Thank u
Is there another simpler method?
@

@9xLn7XtCQi4 years ago Can we please have access to the files with the images of road signs, so we can download it. Thank you. 7
@

@subikshaks42834 years ago Awesome! Easy to understand. One request, can I get the link to the datasets? I have tried all the links in the comment section(which you have provided). Either it says Error 404 or it is unable to download.
@

@yoshirovilchez33922 years ago Which specifications would be the best for the cam ?
@

@mr_unknown_x132 years ago Hi Murtaza I need your help that project traffic sign classification CNN project which version of keras library would rather to use because of that library does not fix my computer architecture and one problem is in library keras of object Sequential there is no get_distribution_strategy argument please help me What should I do ? ...
@

@user-di1wu3oh3o4 years ago Good day! Could you tell me, please, am I able to compile "TrafficSign_" to "*.exe" file using pyinstaller? Thanks! 1
@

@nirbhaychaudhary83374 years ago i used 142 categories in place of 43 categories and there is an error coming " The dimesions of the Training images are wrong
" can u plse help me how to solve it 1
@

@sm02hero20last year Greetings!
Sir I have seen your youtube vedio regarding THE GERMAN TRAFFIC SIGN RECOGNIATION BENCHMARK
which is very interesting vedio I need this project then i have downloaded three things:
1) Labels
2) Data set
3) training source code
But the problem is how I join these three things to run a project please need a guide to run this project to step by step.
Thanks & regards ... 1
@

@emreay95774 years ago Thanks for this very informative and clear lesson. I try exactly the same code but in while loop it trys to recognize everytime and gives predictions even my face front of it. As you guess it's all wrong predictions. Also when i show real signs(from my phone) it also predicts regardless what has been showed as it done the same for my face. Besides when i see the result screen, it's a little bit laggy. Is it because of my computers device feature or i did something wrong? ...
@

@syedumaidahmed62954 years ago Sir ! I really need your advice. I am training a dataset of 52 playing cards with 8,300 images. I want to classify it in real time for AI based card playing game. I am exactly following the tutorial of edge Electronics.
Do you suggest or advice me to try this tutorial of yours ? Please guide.
I have read your comment that this is good only in controlled environment. ...
@

@rafihoque74316 months ago Sir, Can you please provide the versions of the libraries used here? Code is not running due to version issue 1
@

@nicusimanovschi56794 years ago I want to do a change and I can't do it for a few days. I wanna change instead of output text for output image with sign. Every time when i try to do I get error. Can you help me pleas ? How can I put the imagine in main Windows ? ...
@

@tanujak65773 years ago For anyone getting errors despite copying the exact same code, try specifying the version for the packages you install. For keras use 2.3.1and for tensorflow use 2.0.0, like he uses in the next video. ... 16
@

@anshumansharma12843 years ago if i dot hav knowledge about opencv so shall I go for this or should learn opencv before
@

@user-di1wu3oh3o4 years ago Could you tell me please, how can I plot distribution data after augmentation?
@

@jonathannilam66924 years ago Hello sir good day, can you make tutorial for traffic signs classifications using YOLO or MobilenetSSD? Thanks
@

@anuraggupta23834 years ago Sir, from where we will get the dataset? 1
@

@hichamhmidani58752 years ago if I would testing an outher video (like when I driving a car) it is worked in test ?
@

@ShreyasAnupDixit4 years ago Thanks for this video! Can you please tell me the name of the dataset?
@

@lupin84902 years ago Why the camera didn't work with me ? Did I should use an external camera or I can use the that in my laptop?
@

@pakjeromancing27734 years ago Hello, thank you the video, but i have a question, how to stop the train ?
@

@31_kamble_pranav42 years ago Sir pls reply i was working on number detection and sign detection but both are not working efficient after copying your code 1
@

@avirupdey75653 years ago which version of opencv are you using?
@

@imanemanar29613 years ago Hi sir thank you so much for your explanation I followed all of your instructions and it worked but when I tried to run the project in Raspberry Pi 4 it stops in the training and I get (process finished with exit code 135 (interrupted by signal 7: sigemt)) can you please help me with that ? ...
@

@user-hh4du9ry9g2 years ago Is it possible to write the test code using pickle file in C++? Plz answer, thx.
@

@remyaprasad32674 years ago sir can you please give the details of base papers you reffered for this project.
@

@adityakole9854 years ago Can i use laptop's front camera instead of webcam and if yes what changes am i supposed to do in code ? 1
@

@vijayphanindramatha57732 years ago Hello, could you please attach the latest code for traffic sign recognition using camera. I couldnt run this code on my machine
@

@adityaippili67344 years ago Can you state which python version as tensorflow is not compatible with the latest 3.8.1 ??
Pandas is showing an error.
Can you put a picture of all the packages you have installed ?? 1
@

@khanhlinh-mechatronicsengi93883 years ago How can I download the code,please help me, thank you so much 2
@

@jackjacki81914 years ago hello sir thanks for sharing i have some quetions what is the structure used and , number of layers, and types of layers, ...)
and if i want to add the detection what is my step followed ???
@

@hirushkaekanayake72533 years ago how can i get the dataset of this project? its not downloadable which is in your website.
@

@wsterquino3 years ago I havent found the github repository link. Could you please share it? 1
@

@Celesica2 years ago Hello, is it possible to detect the weight and size of an object as well? Thanks!
@

@shafiqurrehman84764 years ago Salam, sir murtaza r u available for paid projects??
@

@thespectre58453 years ago Hello. Can we implement this on pi 4
Great work btw 1
@

@mariamamin20734 years ago which python version you use and also tell the version of tensorflow
@

@dr.ahmedal-adaileh29014 years ago Dear Hassan, many thanks for sharing..
Unfortunately, I did not manage to find the repository neither on your website nor in GitHub.
Is it possible to share the repository link with us?
Cheers
@

@bsungur4 years ago Hello Sir, Thanks for the awesome project. I need your help with the camera issue. I was able to achieve run the project in Rasberry Pi (Raspbian GNU/Linux, VERSION=10). But when I run this project my integrated camera is opening and the camera turns to a white screen. How can I solve this problem? ... 1
@

@TheOtherGuyesShow14 years ago Can I write all these codes in one program? If possible how can I do that? Please suggest me
@

@narotian4 years ago can we add more images to this data and evaluate?
@

@khainguyen56634 years ago Hello everyone, I have a problem when i train a data set in this videos.
WARNING:tensorflow:Your input ran out of data; interrupting training. Make sure that your dataset or generator can generate at least `steps_per_epoch * epochs` batches (in this case, 20000 batches). You may need to use the repeat() function when building your dataset.
If you know how to resolve my problem please comment below ... 5
@

@J11LM4 years ago Thank you for this video.
I'm getting an error after it import all folders (just after the 'data shape (101, 1) <class 'pandas.core.frame.DataFrame'>' line:
File "C:/Users/--/PycharmProjects/Trafficsigns/TrafficSigns/TrafficSign_", line 87, in <module>
axs[j][i].imshow(x_selected[random.randint(0, len(x_selected) - 1), :, :], cmap=plt.get_cmap("gray"))
IndexError: index 43 is out of bounds for axis 0 with size 43
Do you know what can I do to solve it? ...
@

@bsungur4 years ago Dear Murtaza,
The Epoch process is very taking the time. How can I reduce the processing time? Because I need the accuracy - loss graph. Thank you.
(Epoch 1/10 115/2000 [>.............................] - ETA: - loss: 3.4694 - accuracy: 0.0821) ...
@

@anilnaikan2634 years ago sir i need this
code and file plz help me out in this
and sir will it work in any python and pycharm version
like all libraries like keras, tensarflow, pil, pandas i am facing issue in installing all this libraries plz help me out in this sir ...
@

@simpsonchiwashira55124 years ago hi, i am doing an autonomous vehicle with neural network project. However i am using raspberry pi and arduino for this project, and i also want it to detect traffic lights, if i have my dataset for traffic lights can i include it and what changes would i have to do? Please HELP! ... 1
@

@hichamsekkil52603 years ago Is the link to download the dataset available, please?
@

@tueoanminh99765 months ago Hi! can i have the pre-trained model? i copy the same code but it's not working for me, there are some errors...
@

@robotix16694 years ago hello, can we run this in RPI .because it is using TensorFlow Thank you
@

@mzkhan40233 years ago Awsoom video Can I enroll for this course traffic sign board detection pz share me link
@

@anjalik65243 years ago Sir the code for this is not available in the above mentioned link in description.. kindly guide me
@

@IsmaelRivasHidalgo4 years ago Great video! I execuded the same code with the same dataset but I got an error on Train file:
pickle.dump(model, pickle_out)
TypeError: cannot pickle 'weakref' object
Can you help me, please? 1
@

@MePeterNicholls3 years ago My car often gets 50 and 60 mixed up. I think I’m the uk the 5 can look very close to 6 on our road signs, but when it makes mistakes it’s not obvious why
@

@humoruniverse63993 years ago Dude, I wanna create a license plate classification like this. Do you have any examples of it?
@

@magapachanov5157last year Hi where to get the folder model_rtained.p please can you give could not find in the links
@

@Wanderlust13423 years ago what to do with data imbalance such as your is there any way to balance all the classes
@

@buntybois4 years ago Sir, I'm getting assertion error, the dimensions of Training images are wrong
I've kept my imageDimensions as (32,32,3) 1
@

@MrUnbeatable-zz3ve3 years ago Getting error like this while training:
tensorflow:Your input ran out of data; interrupting training. Make sure that your dataset or generator can generate at least `steps_per_epoch * epochs` batches (in this case, 20000 batches). You may need to use the repeat() function when building your dataset.
How resolve this? ...
@

@rajarshibhattacharya2584one week ago How you use keras... Gives me error in keras.model and other packages.... Also I installed all the packages correctly... List the packages pls 1
@

@danrusuciuc88183 years ago Hi Murtaza, can you help me with the issue that appears when i try to save the model
Traceback (most recent call last):
File "C:\Users\Dan\PycharmProjects\Model_rec\", line 215, in <module>
pickle.dump(model, pickle_out)
TypeError: cannot pickle 'weakref' object
what means this weakref object?
Thank you, amazing video btw!! ...
@

@damienguillon25524 years ago Hello, I get the following output: Traceback (most recent call last):
File "C:/Users/damie/PycharmProjects/pythonProject/", line 214, in <module>
pickle.dump(model, pickle_out)
TypeError: cannot pickle '_thread.RLock' object
This is for training the data ... 1
@

@mohanm27182 years ago please tell where is it the pickel file "model_trained.p"
@

@apoorva_n3 years ago The system cannot find the path specified:'myData'..why I am getting this error even after having my data file
@

@pavankanumuri20883 years ago Can u please tell place for code and dataset description link not working
Please
@

@jonelhong73123 years ago How am I going to resolve a problem saying that memory error: unable to allocate 20.6 gib for an array