This guy deserves your money. He is saving us hours and hours of time by telling us the pitfalls and such. Hes working for us--thanks Paul McWhorter! Also--Looks like this version in 2020 shows '4.1.1'.
Thank You!! I never would have gotten this far into this without your instruction (nano lesson 10). Your style of presentation is TOP- NOTCH. I have not seen anything better in 77 years.
I would like to say here that your lessons on Arduino really helped me on a Module titled Embedded System in my Uni. I had 98% on the module. I'm really grateful to come across your lessons then. I am building up myself ready for another module on AI this coming academic session.
Hello Paul, after 8 installations of the Jetpack it works, I got the OpenCV 3.3.1. The only way is to run the Jetpack version 4.2.3 and is very important to not start the Jetson Nano with the network cable connected. So the steps are: 1 download the Jetpack 4.2.3 from the Nvidias Jetpack archive (not the latest) 2 follow your steps as in lesson 1 to copy the package in the SD card. 3 insert the SD card 4 connect a monitor to the HDMI connector, not the Display port (other way you do not see the start) 5 do NOT connect the network cable ( other ways it will try to update at the same time as you start it the first time) 6 When it start it will ask to upgrade the Ubunto version to the latest, DO NOT UPGRADE! it takes several hours and it update the python to the latest version, and you do not want it. 7 open the terminal and run sudo apt-get update, it will take about 20 min. unfortunately it update from Python 3.6.8 to 3.6.9. 8 run the command sudo apt-get install python3-opencv, it will give you an error and it says the you have to run a pkg manually, just copy the text with the instructions and run it, 9 now can you run sudo apt-get install python3-opencv and it install Open CV version 3.2.0 10 unistall it with the comand sudo apt-get remove python3-opencv and now you have OpenCV 3.3.1. 11 Done !
I'm working overtime. I got OpenCV 3.3.1 installed correctly. No problems! Thank you for sharing all these secrets! This would have been a nightmare to try to figure out on my own. I'm about ready to tackle lesson 11 and I'll be all caught up. I going to bed right now though.
I was able to follow along really well. For each example I cut and pasted each one into the file so there are several commands to make each graph seperately. I also figured out how to install pylint. If you would like the instructions please let me know. Thanks
Dear Paul, I just followed the step you advised in 9:18 to obtain the '3.3.1' Cv2 version, but when I check at 10:28 it appears the version '4.1.1'. After, I worked with this version in the next lesson (11), but the image in video size (WebCam Logitech) is quite big, and latency is high. Please, do you have some suggestion?. Many thanks in advance! Lorenzo
Trying this a bit later than originally launched. Your trick for uninstalling opencv is actually upgrading from 3.2.0 to 4.1.1 ; I will proceed with the lesson in opencv 4.1.1 and hopefully there are no issues. Otherwise I will report back. Thanks!
Perhaps you were on Jetpack 4.3 or 4.4. The lessons were made on Jetpack 4.2. Things can be slightly different. I dont think opencv 4.1.1 will be a problem.
I have the same situation. Did you run into any issues. I am struggling getting the examples in lesson 12 to work, in particular the the convert to greyVideo.
In the first lesson Paul explains how to install JetPack 4.2.2. However, since then the JetPack versions have been upgraded and will not always work with his lessons. In order to follow along with him and assure the lessons all work, go to developer.nvidia.com/jetpack-422-archive and follow the install instructions for JetPack 4.2.2.
I'll take a wild guess and say 4.1.1 will work its the updated model so it updates to the latest of everything this thing is super fast I haven't gained this much knowledge since tutorials 3 4 lol awesome videos Paul
Hi Please on following lesson 10 installing OpenCV for Python3, cv2 version came up to 4.1.1 Would this be a problem on following after the rest of the videos?
For these videos I was on Jetpack 4.2, and most of you guys now are on Jetpack 4.5. There will be a few snags along the way, but if you hit one, usually you will find helps in the comments. With a few tweaks everything should work. Good luck
Hi, I am in catching up mode here. Great course! Here are some late comer notes: 1) The Jetson Nano version that you are using (A02?) has been replaced with version B01. As a result, I am using the B01. B01 and I are through lesson 9 and so far so good. 2) The Yahboom Acrylic Case for NVIDIA Jetson Nano A02 is being updated to work with B01. These should be available June 15th, 2020 or so. 3) Is there a way to get both version B01 Pi Cameras working? - Jim
Late reply, but here are the answers. 1. Yes, it will work totally fine. Paul IS using the A02 as far as I can see. 2. Yay! 3. Yes - tutorial 47 explains this.
This is great, but a solution to do this headless (remote ssh into Jetson) would be so much more convenient. VNC seems not to work. Thanks for this excellent series.
There are ways to make vnc work, but do not like editing system files. I have made it work, but on my main system will wait for a more stable solution. Missed you live this morning.
Hey sir, can you please do a tutorial on IoT using Firebase with esp32? It will help a lot of people. There are currently no good in depth tutorials/articles on this device.
Hi Paul, On my Python 2.7 version, there is no OpenCV installed. On Python 3, OpenCV 3.2.x is installed. I followed your procedure but it keeps installing OpenCV 3.2.x only on Python3. Please help
i have a problem whit my jetson nano , in the terminal show "the version is not the correct" after the sudo apt-get install python3-open cv command , caan you help me mr Paul , amazing videos , good vibes frrom Mexico
All these lessons are built on jetpack 4.2 . You are probably on jetpack 4.4 so started with a more recent openCV. Most things should work with the more recent openCV. One problem will be with the countours command in a later lesson but if you look at the comments for that lesson it will show you how to tweak the contrours command
On lesson 18 currently but it’s at the point where I need the intellisense to work to keep up with all the commands, I’m going through this again and for some reason I can’t get intellisense to autocomplete functions for cv2...everything else works except cv2
Just started out and am now up to Lesson 10. I followed the process of installing OpenCV but ended up with version 4.1.1. I am plunging on in the hope that it is backwards compatible with OpenCV version 3.3.1. I doubt if I am smart enough to somehow get 3.3.1 installed without a great deal of difficulty. I will let you all know how it goes.
I'm not able to get Intellisense for CV2 from within VS Code, after following these directions. I think opencv needs to be PIP installed so it resides with PYLINT. If anybody has a work-around please let me know.
Something even crazier happened to me... After installing and just like you have shown, I´ve got version 3.2.0, then after removing I´ve got version 3.3.1 so far so good, but then I reinstalled again and got version 3.2.0 as expected... The crazy thing is that I removed the package for second time and it was upgraded to 4.1.1
OK, if you flash jetpack 4.3 you do not have to do any of this stuff. It will vome with 4.1.1 without doing anything. The thing in this video was needed for jetpack 4.2. I am sorry for the confusion, but it makes it hard when they change things after I produce the videos.
Hi @@paulmcwhorter I had the same thing happen to me, i removed the package and opencv was upgraded to 4.1.1 based on the comments on this channel, i take it i dont need to do anything else and just continue on with your tutorials?
Hi Paul, After following all your instruction carefully, I'm unable to install cv2. version 3.3.1. I'm getting version 4.1.1. and this version won't work in the python3 for this lesson. Is there any remedy for this issue that you know of lately?
The problem is that this lesson was made for jetpack 4.2 and you are probably on jetpack 4.4. On jetpack 4.4, it comes with openCV 4.1.1. You dont need to upgrade as shown in this lesson if you are already on openCV 4.1.1. That should work the same for most of the lessons. I am not sure why you are posting that it does not work. I know a number of people using these lessons on openCV 4.1.1 and most things are identical.
I don t know if you can help me out doctor but i am using arducam UC-517 motorized focus , imx477 , 12 megapixel HQ , the problem i am having i am getting -121 error. ... And i am stuck here .
I'm watching this as of Aug 2020 & have followed the load/ unload procedure & its installed OpenCV v.4.1.1 (as opposed to v.3.3.1). I just wondered if this is going to be an issue moving forward with these videos? I'm wondering if I should try an install an older version??
On jetpack 4.3 or 4.4 you do not need to do this step, you already have the recent OpenCV. Most things in these lessons should work on the openCV you have
@@paulmcwhorter I was almost going to check, before I went & followed the procedure (but got carried away & didn't), so its now running the v4.1.1. I'm just on the next lesson setting up the Camera (im using the camera that came with the Jetbot - I see it has the same Sony IMX219 sensor as the Pi 2 cam & seems to be working OK - code line 5: Flip = 4 (not :2) to display the image the right way up though.
Hey guys, so a little update, 4.1.1 seems to work but i was following along in lesson 11 and ran in to an error, File "/home/ctcombs0351/Desktop/pyPro/openCV/openCV2.py", line 13, in outVid.write(frame) AttributeError: 'tuple' object has no attribute 'write' It also didnt seem to be recognizing the (VideoWriter) command, the auto fill wasnt picking it up. it was'nt creating an error but idk. outVid=cv2.VideoWriter,('videos/myCam.avi',cv2.VideoWriter_fourcc(*'XVID'),21,(dispW,dispH)) Any thoughts on whether this is an issue with openCV 4.1.1 or possibly something else that hasnt installed correctly, or have i made a painfully obvious error in typing in the commands? -THANKS
being able to download and install software just by the command in the terminal, makes me wonder if there's a computer virus concern... Not necessarily this opencv, but some other app.
Hi Paul, love your series I'm digging diligently through it. I'm not sure why perhaps because I'm running Python 3.6.9 but when I go to remove CV2 it removes it. and then kicks me in error" no such module" anytime you could help me with this I would surely appreciate it. I have tried installing CV to 3.3.1 through alternative means and have failed totally.
If you start with jetpack 4.3, which you might have, you do not need to do this step. When I made these videos we were on jetpcak 4.2, and had to do these steps.
I just got a NANO and have been itching to try it out. I am running through this lesson to get caught up and the open CV version I have installed now is 4.1.1 I havent tried it yet but I will let you all know if it works!
I think you probably are on a different version of jetpack. They released a newer version after I made the videos. I dont know what will work, or what will not, as there are so many different possiblities. I would probably suggest going back and using the earlier version that was official release on the date I made lesson 1.
A possible alternative way to repair the OpenCV supplied on the 2019/07/19 SD card image. When researching OpenCV problems I was having using the 2019/12/17 SD card image, I switched to the 2019/07/19 image, the one Paul is using. Having hindsight and foresight now, and starting with a fresh card image, I decided to look more closely at what is wrong with the OpenCV supplied on the 2019/07/19 image. I posted information about my OpenCV problems in Lesson 11. Within that post is a link to a post I made to the nvidia help forum: devtalk.nvidia.com/default/topic/1070685/jetson-nano/jetson-nano-latest-sd-card-image-r32-3-1-contains-opencv-4-1-1-but-no-quot-py-quot-source/ Look at the update I made 02/05/2020. In it are steps I followed using a fresh 2019/07/19 SD card image. Not stated there, I was following what has been done in the videos here. Since my SD card image was fresh, I had not yet installed matplotlib. Step 3 there was a sanity test of OpenCV and python3, done well in advance of Lessons 9-11.. You will see that, indeed, OpenCV is not (properly) installed. However, there is a clue -- the "import cv2" fails due to a missing numpy.core.multiarray. Further down, the "import numpy as np" fails due to missing numpy. Just for grins, I decided to install matplotlib (what was done in Lesson 9) then try the cv2, python3 sanity test again. Success! (There is a typo at step 4: I ran "sudo apt-get install python3-matplotlib", not 'sudo apt-get install matplotlib".) I've left my Jetson Nano with this being the fix for a bad OpenCV. No problems so far. Summary. It is possible that OpenCV was "fixed" after Lesson 9, eliminating the need to do the "install, remove" of OpenCV.
Tried installing opencv on Jetson nano 2gb variant with this funky method. Can't get past version 3.2.0 after 5 tries. Also, sudo apt-get install python3-opencv=3.3.1 is also not working
Hi Paul, I just installed the image 3.6.9 from nano web site. The opencv is 4.1.1 after following the removing command. Will this work? If not how can I install the 3.3.1 version of openCV. Thanks in advance.
If you are on jet pack 4.3 you do not need to do the uninstall/install step, you have openCV 4.4.1 and that will be good. A couple of commands will be different, specifically when you get to the contours lesson, the cv2.findcontours will return only two parameters and not three. Most other commands will be the same
With the latest python3 version, I eventually saw this as I copied Paul's instructions and then had cv2 version 4.1 as shown: barry@barry-nano:~$ python3 Python 3.6.9 (default, Apr 18 2020, 01:56:04) [GCC 8.4.0] on linux Type "help", "copyright", "credits" or "license" for more information. >>> import cv2 >>> print(cv2.__version__) 4.1.1 >>>
Me too, although mine says 4.1.1. If you read all the comments below, you can either go back to square one and install an older JetPack or go forward and probably read through the comments to see what people do to make it work. Paul made these videos a year ago and unfortunately there have been a lot of changes. I am going to stick with my 4.1.1 and see how far I get.
Hi Paul - I'm using the new 2GB NANO that was just released. Thanks for the lessons! The PI camera is working but I am getting an error in the terminal when I run the program. Here it is: [ WARN:0] global /home/nvidia/host/build_opencv/nv_opencv/modules/videoio/src/cap_gstreamer.cpp (933) open OpenCV | GStreamer warning: Cannot query video position: status=0, value=-1, duration=-1 It doesn't seem to affect the execution of the program but I don't know how to get rid of it. Has anyone else seen this error?
Hello I come back from the future, lesson 28, where we do the drawing of the contours. I had version 4.1.1 it worked up until that video with some of the functions we used. I made the mistake of removing it and now I am stuck with version 3.2 because this method only works the first time you do it. Does anyone know a different way to install version 3.3.1 or at least upgrade me back to the version I had?
I think I would flash a fresh Jetpack 4.3, then dont mess with OpenCV and you will have the right version. The only thing different is the contours command returns only two parameters so you will need to adjust that line of code
I followed the instructions and after install CV...then removing...the version for OpenCV is now 4.1.1. I hope this won't cause issues moving forward as I'm not sure how to get a specific version.
I am really not sure . . . it could be you are using a more recent NVIDIA image, and perhaps it comes with a more recent OpenCV. I have not used 4.1.1 so do not know how much might have changed.
@@paulmcwhorter I just purchased my unit late last week so it's new. So far it is operating fine thru most of lesson 11. I appreciate your videos. Thank you.
@@paulmcwhorter For some unknown reason I filled up my 128G drive and had to start from scratch installing software. Like Steve, I find myself locked on OpenCV version 4.1.1. Unfortunately I get the following error when working in Lesson 11 :AttributeError: module 'cv2' has no attribute 'waitkey.' So I am stuck here. I have been working on solving the problem for a week without success. If anyone discovers how to solve, I would appreciate a note here. Ted
first of all: Thanks for your great tutorials! I am new to Jetson Nano and appreciate your lessons so much. But I got a Problem with my Version of openCV. I am running Python 3.6.9 and after uninstalling python3-openCV I am left with opencv 4.1.1. I wouldn't bother, but I am getting error messages in VS-Code regarding my imports. None of the functions of cv2 are recognized and are marked as error. e.g. "Module 'cv2' has no 'VideoCapture' member...." . Never the less it is compiling and running, when I insert a shebang that points to python3. How do I get opencv-4.1.1 downgraded to opencv 3.2.0 to avoid these conflicts while following your lessons? Maybe someone can Help, I'm pretty stuck here. Thanks and cheers to all!! :-)
Andreas, most of the lessons should run with openCV 4.1. I think on the lessons with contours, you have to adjust the contour command to receive 2 parameters instead of three, but most things should run properly. The core issue is that these lessons were made on the NVIDIA Jeston operating image Jetpack 4.2. If you downloaded resently you are probably on jetpack 4.4, and hence that is why you see the differences. You can go back, and from the NVIDIA jetpack archive, install jetpack 4.2, and then everything should work exactly like my lessons. However, if you bravely forge ahead with your existing system, most things actually should work.
@@paulmcwhorter Thanx for the quick reply! I guess I will follow along with my actual system as long as it is running the python files. But I will flash a sd card with Jetpack 4.2 just in case.. I am missing the conveniance of code-prediction in VS-Code.. Big thumbs up for your work! Greetz! :-)
These lessons were for Jetpack 4.2. YOu are probably on 4.5 or 4.4. This step is not needed on the more recent jetpack because you already have what you need.
Hi, thank you for you reply!. I have problem in lesson 13( the last part) so I downgraded to Jetpack 4.2 and when I tried to have CV2 version 3.3.1 by uninstalling CV2 in python3 it was just nothing there, your special strange trick did not work for me. Now I am trying to install CV2 with CMake but I still do not success. I will put the information here when I mange to install CV2
@@paulmcwhorter Jetpack 4.3 (released 12/23/19) came with OpenCV 4.1.1. I don't know if moving from 3 to 4 will cause issues but if 4 is there I'd definitely try that first.
The challenge I have is that I have to make the videos based on the Jetpack as it exists at the time. Then, if they change it, it is not practical for me to redo all the videos. The hard thing is things change quickly in high tech world. In this case, it might make sense to go through the videos with the original Jetpack flash I am using.
Too short to have waited a whole week. I would have liked this lesson and the next lesson, #11 combined. Together they would have still come in at about the one hour mark, which is probably the max length you would want per lesson, so it would have made for a much more fulfilling week's wait. And as you seem to be ahead on queuing up the lessons from the fact I can see up to lesson 16 in the side bar, it seems as if that would have been something easily accomplished. While I understand your weekly videos paced for the absolute beginner, and with which I agree, I think we all could have benefited with a little more "meat" with this lesson 10 meal.
This guy deserves your money. He is saving us hours and hours of time by telling us the pitfalls and such. Hes working for us--thanks Paul McWhorter!
Also--Looks like this version in 2020 shows '4.1.1'.
Thank You!! I never would have gotten this far into this without your instruction (nano lesson 10). Your style of presentation is TOP- NOTCH. I have not seen anything better in 77 years.
so easy to follow along with your tutorials, there needs to be more teachers like you around.
I would like to say here that your lessons on Arduino really helped me on a Module titled Embedded System in my Uni. I had 98% on the module. I'm really grateful to come across your lessons then. I am building up myself ready for another module on AI this coming academic session.
Fantastic lesson 10/62. You make it go so easy. Now cv2 is v. 4.1.1 on my Nano. Thanks Paul.
Doing it now, gets 4.1.1! Still true
@@bluluke does 4.1.1 work?
Hello Paul, after 8 installations of the Jetpack it works, I got the OpenCV 3.3.1.
The only way is to run the Jetpack version 4.2.3 and is very important to not start the Jetson Nano with the network cable connected.
So the steps are:
1 download the Jetpack 4.2.3 from the Nvidias Jetpack archive (not the latest)
2 follow your steps as in lesson 1 to copy the package in the SD card.
3 insert the SD card
4 connect a monitor to the HDMI connector, not the Display port (other way you do not see the start)
5 do NOT connect the network cable ( other ways it will try to update at the same time as you start it the first time)
6 When it start it will ask to upgrade the Ubunto version to the latest, DO NOT UPGRADE! it takes several hours and it update the python to the latest version, and you do not want it.
7 open the terminal and run sudo apt-get update, it will take about 20 min. unfortunately it update from Python 3.6.8 to 3.6.9.
8 run the command sudo apt-get install python3-opencv, it will give you an error and it says the you have to run a pkg manually, just copy the text with the instructions and run it,
9 now can you run sudo apt-get install python3-opencv and it install Open CV version 3.2.0
10 unistall it with the comand sudo apt-get remove python3-opencv and now you have OpenCV 3.3.1.
11 Done !
I'm working overtime. I got OpenCV 3.3.1 installed correctly. No problems! Thank you for sharing all these secrets! This would have been a nightmare to try to figure out on my own. I'm about ready to tackle lesson 11 and I'll be all caught up. I going to bed right now though.
God bless you for showing this technique of getting the right version of OpenCV on python3 at 10:29!
I repeat all lessons. I'm ok, let's use the OpenCV2. Thank you Paul!
I was able to follow along really well. For each example I cut and pasted each one into the file so there are several commands to make each graph seperately. I also figured out how to install pylint. If you would like the instructions please let me know. Thanks
Oops this comment should have been put in the previous video (Lesson 9).
Dear Paul, I just followed the step you advised in 9:18 to obtain the '3.3.1' Cv2 version, but when I check at 10:28 it appears the version '4.1.1'. After, I worked with this version in the next lesson (11), but the image in video size (WebCam Logitech) is quite big, and latency is high. Please, do you have some suggestion?. Many thanks in advance! Lorenzo
...(after a while) ... latency and size resolved at Class 13 with cam.set for web cam. Still the version is 4.1.1 . happy!
@@lorenzoleongutierrez7927 Thaks! You are my hero 🙌🙏
Getting closer. Have a good weekend.
getting there during lock-down . . . lots of time to FOCUS . . . Thank You
Enjoyed the "whole enchilada" lol. Great lesson and thanks.
Trying this a bit later than originally launched. Your trick for uninstalling opencv is actually upgrading from 3.2.0 to 4.1.1 ; I will proceed with the lesson in opencv 4.1.1 and hopefully there are no issues. Otherwise I will report back. Thanks!
Perhaps you were on Jetpack 4.3 or 4.4. The lessons were made on Jetpack 4.2. Things can be slightly different. I dont think opencv 4.1.1 will be a problem.
I have the same situation. Did you run into any issues. I am struggling getting the examples in lesson 12 to work, in particular the the convert to greyVideo.
In the first lesson Paul explains how to install JetPack 4.2.2. However, since then the JetPack versions have been upgraded and will not always work with his lessons. In order to follow along with him and assure the lessons all work, go to developer.nvidia.com/jetpack-422-archive and follow the install instructions for JetPack 4.2.2.
Hi paul,
Could you please tell me exactly on which jetson nano you are working on this series it is 4.2 or 4.2.3
Your words are made of gold, this subject is incredible, you make this even more powerful
Amazing work Paul. If I wasn't so buried in debt I would find a way to financially repay you. Maybe at a later date.
Great as always however have a little problem - cv2 is now on version 4.1.1? is this a problem?
I'll take a wild guess and say 4.1.1 will work its the updated model so it updates to the latest of everything this thing is super fast I haven't gained this much knowledge since tutorials 3 4 lol awesome videos Paul
Excellent teacher ❤❤❤.
Love your videos!
Hi Please on following lesson 10 installing OpenCV for Python3, cv2 version came up to 4.1.1 Would this be a problem on following after the rest of the videos?
I'm wondering about the same issue. I'm on Ubuntu 20 and have Python 3.6.9
What have you found out?
Thanks
openCV4.1.1 but based on some of the comments and responses I'm sure it will be just fine..looking forward to the next lesson.
For these videos I was on Jetpack 4.2, and most of you guys now are on Jetpack 4.5. There will be a few snags along the way, but if you hit one, usually you will find helps in the comments. With a few tweaks everything should work. Good luck
Thank you, Paul. how to install open-cv contrib (with extra modules) ?
Sir, What topic is currently trending internationally in Electrical engineering. (Like, we have python, ML in computer field) .
Extraordinary ...full of knowledge..Thank you sir..long live.
Thank You!!!!!!!!!!! for your lessons. they are great.
Glad you think so!
Thanks for this!
Hi,
I am in catching up mode here.
Great course!
Here are some late comer notes:
1) The Jetson Nano version that you are using (A02?) has been replaced with version B01. As a result, I am using the B01. B01 and I are through lesson 9 and so far so good.
2) The Yahboom Acrylic Case for NVIDIA Jetson Nano A02 is being updated to work with B01. These should be available June 15th, 2020 or so.
3) Is there a way to get both version B01 Pi Cameras working?
- Jim
Late reply, but here are the answers.
1. Yes, it will work totally fine. Paul IS using the A02 as far as I can see.
2. Yay!
3. Yes - tutorial 47 explains this.
This is great, but a solution to do this headless (remote ssh into Jetson) would be so much more convenient. VNC seems not to work. Thanks for this excellent series.
There are ways to make vnc work, but do not like editing system files. I have made it work, but on my main system will wait for a more stable solution. Missed you live this morning.
Hey sir, can you please do a tutorial on IoT using Firebase with esp32? It will help a lot of people. There are currently no good in depth tutorials/articles on this device.
Hi Paul,
On my Python 2.7 version, there is no OpenCV installed. On Python 3, OpenCV 3.2.x is installed. I followed your procedure but it keeps installing OpenCV 3.2.x only on Python3. Please help
i have a problem whit my jetson nano , in the terminal show "the version is not the correct" after the sudo apt-get install python3-open cv command , caan you help me mr Paul , amazing videos , good vibes frrom Mexico
All these lessons are built on jetpack 4.2 . You are probably on jetpack 4.4 so started with a more recent openCV. Most things should work with the more recent openCV. One problem will be with the countours command in a later lesson but if you look at the comments for that lesson it will show you how to tweak the contrours command
Love it!!!
On lesson 18 currently but it’s at the point where I need the intellisense to work to keep up with all the commands, I’m going through this again and for some reason I can’t get intellisense to autocomplete functions for cv2...everything else works except cv2
Just started out and am now up to Lesson 10. I followed the process of installing OpenCV but ended up with version 4.1.1.
I am plunging on in the hope that it is backwards compatible with OpenCV version 3.3.1.
I doubt if I am smart enough to somehow get 3.3.1 installed without a great deal of difficulty.
I will let you all know how it goes.
Did the different Opencv version give you any trouble John? Im in the same boat
Following this process I ended with OpenCV version 4.1.1. on 12/11/20. I hope this version will be OK.
Same here, getting worried.
I'm not able to get Intellisense for CV2 from within VS Code, after following these directions. I think opencv needs to be PIP installed so it resides with PYLINT. If anybody has a work-around please let me know.
I'm about to make a big order from Amazon. Is there anything else I need for later in this series? What about the robot series?
A Gaggia One Touch Cappuccino Machine.
@@paulmcwhorter I'll get right on that!! 😁
Something even crazier happened to me... After installing and just like you have shown, I´ve got version 3.2.0, then after removing I´ve got version 3.3.1 so far so good, but then I reinstalled again and got version 3.2.0 as expected... The crazy thing is that I removed the package for second time and it was upgraded to 4.1.1
OK, if you flash jetpack 4.3 you do not have to do any of this stuff. It will vome with 4.1.1 without doing anything. The thing in this video was needed for jetpack 4.2. I am sorry for the confusion, but it makes it hard when they change things after I produce the videos.
Hi @@paulmcwhorter I had the same thing happen to me, i removed the package and opencv was upgraded to 4.1.1 based on the comments on this channel, i take it i dont need to do anything else and just continue on with your tutorials?
On JetPack 4.6 this just removed my openCV :(
Hi Paul, After following all your instruction carefully, I'm unable to install cv2. version 3.3.1. I'm getting version 4.1.1. and this version won't work in the python3 for this lesson. Is there any remedy for this issue that you know of lately?
The problem is that this lesson was made for jetpack 4.2 and you are probably on jetpack 4.4. On jetpack 4.4, it comes with openCV 4.1.1. You dont need to upgrade as shown in this lesson if you are already on openCV 4.1.1. That should work the same for most of the lessons. I am not sure why you are posting that it does not work. I know a number of people using these lessons on openCV 4.1.1 and most things are identical.
for me it appear version open CV 4.1.1 it says it is compatible with python3 3.6
Thank you for the video! is it possible to make it work inside a virtualenv? Thanks
china, aliexpress
I don t know if you can help me out doctor but i am using arducam UC-517 motorized focus , imx477 , 12 megapixel HQ , the problem i am having i am getting -121 error. ... And i am stuck here .
Hey Paul quick question I’m having problems using ir sensors in daylight, is there a way of being able to use ir sensors efficiently in sunlight?
You'll want to put a shroud on them, even just sliding a small tube over them will usually suffice.
4:03 - It might also be helpful to consider using a class A1 SD card.
I'm still getting the 3.2.0 no matter what I do I did 3 times and still the same is there another way
I'm watching this as of Aug 2020 & have followed the load/ unload procedure & its installed OpenCV v.4.1.1 (as opposed to v.3.3.1). I just wondered if this is going to be an issue moving forward with these videos? I'm wondering if I should try an install an older version??
On jetpack 4.3 or 4.4 you do not need to do this step, you already have the recent OpenCV. Most things in these lessons should work on the openCV you have
@@paulmcwhorter I was almost going to check, before I went & followed the procedure (but got carried away & didn't), so its now running the v4.1.1. I'm just on the next lesson setting up the Camera (im using the camera that came with the Jetbot - I see it has the same Sony IMX219 sensor as the Pi 2 cam & seems to be working OK - code line 5: Flip = 4 (not :2) to display the image the right way up though.
Hey guys, so a little update, 4.1.1 seems to work but i was following along in lesson 11 and ran in to an error,
File "/home/ctcombs0351/Desktop/pyPro/openCV/openCV2.py", line 13, in
outVid.write(frame)
AttributeError: 'tuple' object has no attribute 'write'
It also didnt seem to be recognizing the (VideoWriter) command, the auto fill wasnt picking it up. it was'nt creating an error but idk.
outVid=cv2.VideoWriter,('videos/myCam.avi',cv2.VideoWriter_fourcc(*'XVID'),21,(dispW,dispH))
Any thoughts on whether this is an issue with openCV 4.1.1 or possibly something else that hasnt installed correctly, or have i made a painfully obvious error in typing in the commands? -THANKS
I have python 3.6.0 When I type import cv2 enter I get Illegal instruction ( core dumped ) . Does anyone know why ?
being able to download and install software just by the command in the terminal, makes me wonder if there's a computer virus concern... Not necessarily this opencv, but some other app.
Hey my python 2 has version 4.4.1
Please help me
Is amazing, si thanks for this video.
Hi Paul, love your series I'm digging diligently through it. I'm not sure why perhaps because I'm running Python 3.6.9 but when I go to remove CV2 it removes it. and then kicks me in error" no such module" anytime you could help me with this I would surely appreciate it. I have tried installing CV to 3.3.1 through alternative means and have failed totally.
If you start with jetpack 4.3, which you might have, you do not need to do this step. When I made these videos we were on jetpcak 4.2, and had to do these steps.
@@paulmcwhorter Your prompt response speaks strongly of your level of commitment and integrity much appreciated Jim Peterson
I just got a NANO and have been itching to try it out. I am running through this lesson to get caught up and the open CV version I have installed now is 4.1.1 I havent tried it yet but I will let you all know if it works!
version 4.1.1, doesn't seem to work and I'm stuck. I ask Paul for a remedy but haven't heard back yet. Have you found a solution?
After uninstall OpenCV I got: no module named 'cv2'.....
Did you solve this problem? I got this problem too.
Yep, sure did get latest version of openCV, unfortunately it's 4.4.1, so I'm hoping that won't cause grief.
Hi you are grate!
i get version 4.1.1 is that a problem ?
Yes, it seems to be a problem, and doesn't work. Have you solve this yet?
Awesome!
After installing and removing I ended up with cv2 version 4.1.1 NOT version 3.3.1 Will This Work?
I think you probably are on a different version of jetpack. They released a newer version after I made the videos. I dont know what will work, or what will not, as there are so many different possiblities. I would probably suggest going back and using the earlier version that was official release on the date I made lesson 1.
A possible alternative way to repair the OpenCV supplied on the 2019/07/19 SD card image.
When researching OpenCV problems I was having using the 2019/12/17 SD card image, I switched to the 2019/07/19 image, the one Paul is using. Having hindsight and foresight now, and starting with a fresh card image, I decided to look more closely at what is wrong with the OpenCV supplied on the 2019/07/19 image. I posted information about my OpenCV problems in Lesson 11. Within that post is a link to a post I made to the nvidia help forum:
devtalk.nvidia.com/default/topic/1070685/jetson-nano/jetson-nano-latest-sd-card-image-r32-3-1-contains-opencv-4-1-1-but-no-quot-py-quot-source/
Look at the update I made 02/05/2020. In it are steps I followed using a fresh 2019/07/19 SD card image. Not stated there, I was following what has been done in the videos here. Since my SD card image was fresh, I had not yet installed matplotlib. Step 3 there was a sanity test of OpenCV and python3, done well in advance of Lessons 9-11.. You will see that, indeed, OpenCV is not (properly) installed. However, there is a clue -- the "import cv2" fails due to a missing numpy.core.multiarray. Further down, the "import numpy as np" fails due to missing numpy. Just for grins, I decided to install matplotlib (what was done in Lesson 9) then try the cv2, python3 sanity test again. Success! (There is a typo at step 4: I ran "sudo apt-get install python3-matplotlib", not 'sudo apt-get install matplotlib".) I've left my Jetson Nano with this being the fix for a bad OpenCV. No problems so far.
Summary.
It is possible that OpenCV was "fixed" after Lesson 9, eliminating the need to do the "install, remove" of OpenCV.
OpenCV2 is now at version 4.1.1.
Tried installing opencv on Jetson nano 2gb variant with this funky method. Can't get past version 3.2.0 after 5 tries. Also, sudo apt-get install python3-opencv=3.3.1 is also not working
Hi Paul, I just installed the image 3.6.9 from nano web site. The opencv is 4.1.1 after following the removing command. Will this work? If not how can I install the 3.3.1 version of openCV. Thanks in advance.
If you are on jet pack 4.3 you do not need to do the uninstall/install step, you have openCV 4.4.1 and that will be good.
A couple of commands will be different, specifically when you get to the contours lesson, the cv2.findcontours will return only two parameters and not three. Most other commands will be the same
Thanks Paul, I'm able to run openCV.py using webcam but the frame rate is low. How to remedy it. thanks in advance.
I found this : nanocamera pypi.org/project/nanocamera/
It's so easy to set the camera fps. Fixed.
When I remove the opencv from python3 it dosnt replace it with the cv2 3.3.1.
Can I get help?
I would start with a fresh SD card and follow through the video again. it should work.
Why OpenCv instead of using tensorflow/Keras CNN? I assume Keras can run on a Jetson Nano.
Because OpenCV is easy to learn, easy to use, and flexible.
how to take pictures in python with jetpack opencv
thank you.
With the latest python3 version, I eventually saw this as I copied Paul's instructions and then had cv2 version 4.1 as shown:
barry@barry-nano:~$ python3
Python 3.6.9 (default, Apr 18 2020, 01:56:04)
[GCC 8.4.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import cv2
>>> print(cv2.__version__)
4.1.1
>>>
Me too, although mine says 4.1.1. If you read all the comments below, you can either go back to square one and install an older JetPack or go forward and probably read through the comments to see what people do to make it work. Paul made these videos a year ago and unfortunately there have been a lot of changes. I am going to stick with my 4.1.1 and see how far I get.
Hi Paul - I'm using the new 2GB NANO that was just released. Thanks for the lessons! The PI camera is working but I am getting an error in the terminal when I run the program. Here it is: [ WARN:0] global /home/nvidia/host/build_opencv/nv_opencv/modules/videoio/src/cap_gstreamer.cpp (933) open OpenCV | GStreamer warning: Cannot query video position: status=0, value=-1, duration=-1 It doesn't seem to affect the execution of the program but I don't know how to get rid of it. Has anyone else seen this error?
Hello I come back from the future, lesson 28, where we do the drawing of the contours. I had version 4.1.1 it worked up until that video with some of the functions we used. I made the mistake of removing it and now I am stuck with version 3.2 because this method only works the first time you do it. Does anyone know a different way to install version 3.3.1 or at least upgrade me back to the version I had?
I think I would flash a fresh Jetpack 4.3, then dont mess with OpenCV and you will have the right version. The only thing different is the contours command returns only two parameters so you will need to adjust that line of code
@@paulmcwhorter Yea that was the best thing to do thank you
I did the install/uninstall of openCV and now I have 4.1.1 lol
Just got the nano BO1 so looks like its the newer version. Also numpy came installed.
I followed the instructions and after install CV...then removing...the version for OpenCV is now 4.1.1. I hope this won't cause issues moving forward as I'm not sure how to get a specific version.
I am really not sure . . . it could be you are using a more recent NVIDIA image, and perhaps it comes with a more recent OpenCV. I have not used 4.1.1 so do not know how much might have changed.
@@paulmcwhorter I just purchased my unit late last week so it's new. So far it is operating fine thru most of lesson 11. I appreciate your videos. Thank you.
@@paulmcwhorter For some unknown reason I filled up my 128G drive and had to start from scratch installing software. Like Steve, I find myself locked on OpenCV version 4.1.1. Unfortunately I get the following error when working in Lesson 11 :AttributeError: module 'cv2' has no attribute 'waitkey.' So I am stuck here. I have been working on solving the problem for a week without success. If anyone discovers how to solve, I would appreciate a note here. Ted
Sorry but my problem was incorrect spelling of waitKey. I was all lower case. Everything working fine now. Starting to move through the lessons again.
Thank you
first of all: Thanks for your great tutorials! I am new to Jetson Nano and appreciate your lessons so much. But I got a Problem with my Version of openCV. I am running Python 3.6.9 and after uninstalling python3-openCV I am left with opencv 4.1.1. I wouldn't bother, but I am getting error messages in VS-Code regarding my imports. None of the functions of cv2 are recognized and are marked as error. e.g. "Module 'cv2' has no 'VideoCapture' member...." . Never the less it is compiling and running, when I insert a shebang that points to python3. How do I get opencv-4.1.1 downgraded to opencv 3.2.0 to avoid these conflicts while following your lessons? Maybe someone can Help, I'm pretty stuck here. Thanks and cheers to all!! :-)
Andreas, most of the lessons should run with openCV 4.1. I think on the lessons with contours, you have to adjust the contour command to receive 2 parameters instead of three, but most things should run properly. The core issue is that these lessons were made on the NVIDIA Jeston operating image Jetpack 4.2. If you downloaded resently you are probably on jetpack 4.4, and hence that is why you see the differences. You can go back, and from the NVIDIA jetpack archive, install jetpack 4.2, and then everything should work exactly like my lessons. However, if you bravely forge ahead with your existing system, most things actually should work.
@@paulmcwhorter Thanx for the quick reply! I guess I will follow along with my actual system as long as it is running the python files. But I will flash a sd card with Jetpack 4.2 just in case.. I am missing the conveniance of code-prediction in VS-Code.. Big thumbs up for your work! Greetz! :-)
the uninstall trick does not work any more
These lessons were for Jetpack 4.2. YOu are probably on 4.5 or 4.4. This step is not needed on the more recent jetpack because you already have what you need.
Hi, thank you for you reply!.
I have problem in lesson 13( the last part) so I downgraded to Jetpack 4.2 and when I tried to have CV2 version 3.3.1 by uninstalling CV2 in python3 it was just nothing there, your special strange trick did not work for me. Now I am trying to install CV2 with CMake but I still do not success. I will put the information here when I mange to install CV2
Hi again, Jetpack 4.2 did not have OpenCV 3.3.1 but now I am testing with Jetpack 4.2.1 according to the Nvidia blog it shoud have CV 3.3.1
June 5th, 2021. Installed OpenCV, gets version 3.2.0. Uninstall OpenCV, get version 4.1.1. LOL! Go home apt, you're drunk!
The latest Jetpack image (4.3) has OpenCV preinstalled.
Yes . . . which version?
@@paulmcwhorter Jetpack 4.3 (released 12/23/19) came with OpenCV 4.1.1. I don't know if moving from 3 to 4 will cause issues but if 4 is there I'd definitely try that first.
The challenge I have is that I have to make the videos based on the Jetpack as it exists at the time. Then, if they change it, it is not practical for me to redo all the videos. The hard thing is things change quickly in high tech world. In this case, it might make sense to go through the videos with the original Jetpack flash I am using.
@@paulmcwhorter The new Jetpack only has 2 significant updates, OpenCV 4 and TensorRT 6.
what in the world is happening lol how did removing cv2 magically give u cv 3.3.1 stop blowing my mind
thumb up!
Nice hack !
I cant find a fan for jetson in turkey :D if there is a turkish person, can they help me?
search for 4010 5v fan
Yanlış hatırlamıyorsam bir videoda lattepanda fanının uyumlu olduğunu görmüştüm. Ölçülerini kontrol edebilirsiniz.
I have a spare fan in IZMIR
Too short to have waited a whole week. I would have liked this lesson and the next lesson, #11 combined. Together they would have still come in at about the one hour mark, which is probably the max length you would want per lesson, so it would have made for a much more fulfilling week's wait. And as you seem to be ahead on queuing up the lessons from the fact I can see up to lesson 16 in the side bar, it seems as if that would have been something easily accomplished. While I understand your weekly videos paced for the absolute beginner, and with which I agree, I think we all could have benefited with a little more "meat" with this lesson 10 meal.