the training was prepared very well, thank you. I have a question for you. How can I use the weights we trained on Colab in this structure that you have prepared?
Worked first time - that's quite rare for tutorials these days! Well done. I note too that there are lines in the files referring to v4, does the v4 version go faster? Nothing to do with your instructions, but there is a huge lag on the setup I got from this tutorial using a pi-camera so although the detection rate is about 1.5 per second, there is about 2 to 3 seconds delay between reality and the image on the screen. But thanks - this is great!
Thanks, yes, for some reason I also receive lag on the pi, but didnt had time to investigate this problem. V4 is a little more accurate but a little slower, to use v3 or v4 depends on your project goals, on github you can see my v3 vs v4 comparison
Can you show how to use this in conjunction with a NCS2, or point me in the right direction? I have installed open vino, but not sure what im doing wrong.
10.40 the video code and the text side code slightly different and both isnot working in my pi..unable to locate package libgfortran5 please reply soon..
I received error “E: Unable to locate package libatlas3-base” when trying to install the packages, is there any fix for this? I am using Raspberry Pi 3B+
Great video. I am trying to send webcam video over my local network from the pi to a PC which should perform object detection and send the output stream back to the pi over the local network. I hope this will increase fps even after networking overhead. Do you have any tips or experience on how to implement that?
If local network use sockets, fastest way to do so. Send already resized image, try to find what is the most efficient data format to send images through socket. Was not trying to do so, but should work
Sir, I have used your code before, but now I use it. Why is the execution speed slow? When I resize the image, the image is the same size as before during execution i before run fps 1.4 now 0.4 with yolo tiny
Hi, Thank you for you tutorial 🙂 Could you please help me here, While installing tensorflow(last step: sudo -H pip3 install tensorflow-2.2.0-cp37m-linux_armv71.whl) I'm getting an ERROR. Cannot uninstall wrapt. It is a distutils installed project and.... Ps: I'm using pi4 and i did clone yolov4-tiny
sudo pip3 install h5py==2.9.0 SHOWING ERROR: note: This error originates from a subprocess, and is likely not a problem with pip. ERROR: Failed building wheel for numpy Failed to build numpy ERROR: Could not build wheels for numpy, which is required to install pyproject.toml-based projects WARNING: You are using pip version 22.0.3; however, version 22.0.4 is available. You should consider upgrading via the '/usr/bin/python3 -m pip install --upgrade pip' command.
Thank you for this wonderful instruction. Is it possible to use opencv version 3.2.0 ? My R-pi version is 4, though it is hard to install opencv 3.4.6.27. Please give me your kind answer.
hello. this is very awesome. just wont to ask of a possibility to detect a person when he about to leave his bicycle in an inappropriate place. for example, if a person and bicycle together are good. but he wants to leave his bicycle alone the camera detects this and sends an alarm. also, can we detect people riding a bicycle and not riding a bicycle and definite between them?
Yes, all of this is available, but everything depends on where you wanna use it. Commercially or this project is for fun, but usually you may use not only object detection, but you would need to combine object detection with classification models or something like that
Hey was this project done with Yolo V3 or Tiny Yolo (Yolo V3 Tiny)? I've been trying to figure out if Yolo (standard model) is capable of running on Raspberry Pi, and I was under the impression that only the Tiny variant can run on a Pi due to the lack of Nvidia Cuda cores. I'm confused because on your Github repo (github.com/pythonlessons/TensorFlow-2.x-YOLOv3#installation), you have installed yolov3 and yolov3-tiny, as well as yolov4 and yolov4-tiny; were all four of these models compatible with your Pi 3B? Thanks!
@@PyLessons Thanks for responding! I followed your tutorial and got V3 and V4 (plus the tiny variants) working on my Raspberry Pi 4. Although when I ran the detection_demo.py script, I was surprised to find that Yolov3 performs better than v4. Any idea why Yolo v4 is actually less accurate (but faster) than Yolo v3? I read about the v4 improvements and it is supposed to be faster and more accurate at object detection.
Hello congratulations on your tutorials How could it detect only one specific object (bottles)? and turn on a led if there is at least one bottle within the image
Hello. It's quite easy, check my object tracking tutorial, how I implemented code to track only persons, try to use similar technique to detect your bottles. At first try to print to terminal something while bottles are detected, then simply replace print line to led blinking fuction.
this is the most helpful tutorial on youtube. Thank you Bro..
Amazing video, simple and clear for someone starting. I wish everybody can make clear videos like these. Well done Brother.
I LOVE YOU.
YOU EXPLAINED IT SO WELL I LEARNED A LOT
hello , i want to include voice feedback to the project could you tell me please how to do ?
Thanks, but I think you can't do that :)
Thank you very much Sir, this tutorial worked perfectly for me. ready for next step :-)
Nice!
Excelent video, thanks!
I get a 404 error for the tensorflow part when using the link I need help to mitigate this on my raspberry pi 3
the training was prepared very well, thank you. I have a question for you. How can I use the weights we trained on Colab in this structure that you have prepared?
Could you find the solution?
which Rasbperry pi model did you use?
is it possible to convert this model into tensorflow lite and run it with google coral usb?
You are my God!
Hallo Python Lessons, can we change the color of the bounding box freely? thanks before!
Worked first time - that's quite rare for tutorials these days! Well done. I note too that there are lines in the files referring to v4, does the v4 version go faster? Nothing to do with your instructions, but there is a huge lag on the setup I got from this tutorial using a pi-camera so although the detection rate is about 1.5 per second, there is about 2 to 3 seconds delay between reality and the image on the screen. But thanks - this is great!
Thanks, yes, for some reason I also receive lag on the pi, but didnt had time to investigate this problem. V4 is a little more accurate but a little slower, to use v3 or v4 depends on your project goals, on github you can see my v3 vs v4 comparison
Amazing video. Are you using a usb cam? Because I have a usb cam in my rasp pi, but idk how to use it instead the raspberry pi camera.
Yes, usb webcam. I cant explain how to use it, google may tell you better how to do that
Can you show us, how to install correctly Yolo 5?, I think its more faster to use in Raspberry PI 4
Its too hard now to start yolo5
Can you show how to use this in conjunction with a NCS2, or point me in the right direction? I have installed open vino, but not sure what im doing wrong.
10.40 the video code and the text side code slightly different and both isnot working in my pi..unable to locate package libgfortran5
please reply soon..
nice video sir thanks
I received error “E: Unable to locate package libatlas3-base” when trying to install the packages, is there any fix for this? I am using Raspberry Pi 3B+
Great video. I am trying to send webcam video over my local network from the pi to a PC which should perform object detection and send the output stream back to the pi over the local network. I hope this will increase fps even after networking overhead. Do you have any tips or experience on how to implement that?
If local network use sockets, fastest way to do so. Send already resized image, try to find what is the most efficient data format to send images through socket. Was not trying to do so, but should work
you are really awesome, thank you..
Great explanation.
But could you explain to me, how could you set the condition to have an output signal when the camera detects an object?
It's as simple as using if-else statement
Can u detect objects on ur monitor
Sir, I have used your code before, but now I use it. Why is the execution speed slow? When I resize the image, the image is the same size as before during execution
i before run fps 1.4 now 0.4 with yolo tiny
because it's not tiny model and not TFlite? It's quite hard for raspberry to handle everything that's obvious
@@PyLessons i run tiny model with tensorflow no tflite
when i wget tensor, i cant copy link here , isnt it?
Hi, Thank you for you tutorial 🙂
Could you please help me here,
While installing tensorflow(last step: sudo -H pip3 install tensorflow-2.2.0-cp37m-linux_armv71.whl)
I'm getting an ERROR. Cannot uninstall wrapt. It is a distutils installed project and....
Ps: I'm using pi4 and i did clone yolov4-tiny
Didn't tried with pi4, I don't have one, not sure if I can help
@ Aqsa Gaffar same problem any solution?
Dear i want to ask a question :
thanks, its very nice
how can detect custom objects
i have some errors when install tensorflow in terminal. please help me
Pi3 or Pi4? Thanks!
Hello sir ty for ur tutorial it was the best one.
Sir plz can u tell how to increase frame per second FPS?
sudo pip3 install h5py==2.9.0
SHOWING ERROR:
note: This error originates from a subprocess, and is likely not a problem with pip.
ERROR: Failed building wheel for numpy
Failed to build numpy
ERROR: Could not build wheels for numpy, which is required to install pyproject.toml-based projects
WARNING: You are using pip version 22.0.3; however, version 22.0.4 is available.
You should consider upgrading via the '/usr/bin/python3 -m pip install --upgrade pip' command.
Help me! I got an error "ValueError: cannot reshape array of size 4482926 into shape (1024,512,3,3)"
What should I do?
Thanx mate...
Your welcome
Hey can you please tell me is it possible to convert the detected object in to speech. That will help as i am trying to this as a final year project.
LOL, the strange question with an obvious answer. If you can detect it to word, for example, person, you just need to run mp4 which says, person...
Thank you for this wonderful instruction. Is it possible to use opencv version 3.2.0 ? My R-pi version is 4, though it is hard to install opencv 3.4.6.27. Please give me your kind answer.
Not sure, you can try
I wonder if can activate the GPU mode in raspberry? I know that Raspberry has one.
Don't know
hello. this is very awesome. just wont to ask of a possibility to detect a person when he about to leave his bicycle in an inappropriate place. for example, if a person and bicycle together are good. but he wants to leave his bicycle alone the camera detects this and sends an alarm. also, can we detect people riding a bicycle and not riding a bicycle and definite between them?
Yes, all of this is available, but everything depends on where you wanna use it. Commercially or this project is for fun, but usually you may use not only object detection, but you would need to combine object detection with classification models or something like that
@@PyLessons thanks for your reply. can I contact you in pythonlessons0@gmail.com?
@@ahmedzaki2668 Yes you can
Python Lessons hello my dear. I have sent you an email yesterday from az4l88@gmail.com.
Please let me know if it reached you?
Thanks
@@ahmedzaki2668 Hi, I will respond you later, for some reason your email went to spam
I am using Raspberry pi 3A+ but every time my algorithm tries to load the tensorflow the Raspberry just remain unchanged for eternity
As can be seen the video stream from the by camera is stopping. actually it is making delays. how can that be solved?
Hey, I think that for raspberry its too hard to handle yolo detection, maybe this can be a reason
Hey was this project done with Yolo V3 or Tiny Yolo (Yolo V3 Tiny)?
I've been trying to figure out if Yolo (standard model) is capable of running on Raspberry Pi, and I was under the impression that only the Tiny variant can run on a Pi due to the lack of Nvidia Cuda cores.
I'm confused because on your Github repo (github.com/pythonlessons/TensorFlow-2.x-YOLOv3#installation), you have installed yolov3 and yolov3-tiny, as well as yolov4 and yolov4-tiny; were all four of these models compatible with your Pi 3B?
Thanks!
Yes, all of them works. Simply Tiny works faster
@@PyLessons Thanks for responding! I followed your tutorial and got V3 and V4 (plus the tiny variants) working on my Raspberry Pi 4.
Although when I ran the detection_demo.py script, I was surprised to find that Yolov3 performs better than v4.
Any idea why Yolo v4 is actually less accurate (but faster) than Yolo v3? I read about the v4 improvements and it is supposed to be faster and more accurate at object detection.
Hello
congratulations on your tutorials
How could it detect only one specific object (bottles)? and turn on a led if there is at least one bottle within the image
Hello. It's quite easy, check my object tracking tutorial, how I implemented code to track only persons, try to use similar technique to detect your bottles. At first try to print to terminal something while bottles are detected, then simply replace print line to led blinking fuction.
@@PyLessons thankyou very much
@@PyLessons
which video? a lot of videos please help me
great video, congratulations i have my model trained with yoloV3 and i want to run it on the raspberry pi4 please can you help me thanks
I don't have pi4, so I was not able to test the differences with pi3, but it shouldn't be that different
Error on wget for tensorflow. Is there a solution for this?
which raspberry you are using
how to install yolov4?
can i do this with my pi 3 ?
I actually did this on pi3
What is your version of Raspberry Pi ???
From Vietnam with love :3
For this tutorial I used Raspberry Pi 3 Model B V1.2
alo ông cũng làm detection object trên rasperri pi hả =))))
@@nhatminhle2900 t làm trên jetson nano ông ơi
@@quanghuytrinh5985 Kết quả như nào v ông, cái jetson chạy dc real time hem?
@@nhatminhle2900 t train model nhận diện linh kiện điện tử, model nhận diện loại mạch, chạy song song mà vẫn real time
no data or 404 , help me
I got 404 error when wget
I got an error when I try to wget the link
Will try it later my self
@@PyLessons no its fine i figured it out
@@fakhrulamin6958 I'm running into a 404 error on that line, how did you getting it working on your end?
@@fakhrulamin6958 how did you figure it out?
👍
testing
Helo
fatal: fetch-pack: invalid index-pack output
I face getting this error.can you tell me the solution?
no idea, you need to google this friend