I've been through these lessons up to 28 a couple times now on my laptop. I rewrote this set every time just to make sure I got it down. I can write the whole thing on my own now and understand what each line does. Just got my jetson nano gear and moving forward from here on and finally moving past lesson 28.
Who else is donating to Paul and his magnificent, freely available Arduino / Jetson / tech hobby video tutorials!? Everyone!!! Good! Three bucks a month is all it takes, five, eight, ten, you're getting a bargain in online education with Paul and his material.
Everything up running well. Was surprised as was thinking a lot has changed since you posted the video. cv2 version ended up being 3.2.0 but works here for Lesson 11. I do have the Nano with two camera connections. Going to try and use the Luxonis Oak-D Camera with Nano. Thanks for all you do Paul. Great stuff.
Not sure if this video is still monitored; but been enjoying the lessons and was following along so well that I actually paused the lessons until my Raspberry Pi camera came - even though I also had a compatible Logitech camera. Thank you for the lesson plan.
The camera was exciting , I love it when you hold your breadth , you look superb. Your videos are super col and they explain more than just copying and pasting. I love em' all
Enjoying all of your course, Arduino, Python, now Nano. However, with the nano, sudo apt-get install python3-opencv the install is of 4.4.1 and it does not cooperate. Any suggestions or direction would be great as I look forward to completing this series. Is there any method to follow for an install of 3.3.1? Thanks
Hi Paul. I just wanted to say that I've watched all 11 lessons thus far and I am really excited about this course. For reasons I won't go into I haven't been able to follow you on the Jetson Nano, so I'll be playing catch-up this week. Hopefully I'll be all caught up by lesson 12. I have the hardware all set up and running fine. I'm also already familiar with Linux, Python, and Numpy so those will be easy for me. This will be my first experience with OpenCV. At the end of this lesson you mentioned being able to recognize a specific person's face. This is actually my ultimate goal with this project. So I'm really looking forward to learning how to achieve that goal. Thanks for creating these wonderful lessons and I'll be signing up on your Patreon site very soon. You definitely deserve to be compensated for your work. Superb instructions to be sure! Thank you so very much!
I'm all caught up. BOOM! Everything works great. Thank you for creating this superb series Paul. I'll be coming over to patreon shortly to support your hard work.
@@paulmcwhorter That's cool. I'm in no rush. I just wanted to get caught up. I'm working with other things that have to do with AI too. I've been working with numpy and artificial neural networks. So this is just adding hardware to experiment with on top of that. I'm building a robot and one of the things I would like to robot to do is to be able to recognize specific individuals. I see you mentioned that possibility here at the end of this video. So I'm really looking forward to learning how to do that. But like I say, no rush. I'm just happy to be caught up.
That is the direction we are going. First, we will recognize things based on cascades, or training that others have already done, so we can get the nano to do things that are cool by download training sets others have done. Those training sets will allow us to do things like recognize cats verses dogs verses bananas and so forth. Then finally, the final challenge will be for us to create training cascades ourself. This will allow us to recognize a specific person. That is a tall order, but where we are going. I think the tough part of getting an AI system going is the training. Anyway, that is what I am working on. Feel a certain amount of pressure as I have to stay ahead of you guys, so that keeps me on my toes.
@@paulmcwhorter "This will allow us to recognize a specific person. That is a tall order, but where we are going" That's cool. I was going to try to do this on my own. So anything you have to offer will be greatly valued. Just setting the system up this far would have been a nightmare without your help. And I'm sure I'll learn much more from your lessons as well. "Feel a certain amount of pressure as I have to stay ahead of you guys, so that keeps me on my toes." No pressure here. I'm extremely patient. I'm in no rush. Just really happy that you chose to do this course. This course is EXACTLY what I was hoping to find on TH-cam, so I couldn't be more pleased with your choice of subject matter. I think (no scratch that), I KNOW that I couldn't find anything as well organized as what you have presented thus far. You've taken us from the bare-bones assembly right up to getting the camera running being sure to preserve the 128 core capability of the Nano. Can't beat that! I owe you big time, and I'll be over to Patreon very shortly to settle up my debt. I might even do that tonight. But if not tonight, very soon to be sure. You deserve to get paid for all the work you've done. And your product (lessons) are very well worth it. No question about it. In fact, I don't even see how you can do so much as I know you are doing other lessons too. I'm even following some of those as well. You're a hard worker Paul. No question about that.
hello Paul thanks for the lessons i am really enjoying this series of lessons. I have always wanted to play with AI. Keep up the awesome video lessons.
Another great lesson. I thought I would have problems - I'm using the later B01 board with two camera ports, my openCV is 4.4.1 and well Boom all up and working no probs. Sad to have just finished the Fusion 360 lesson set but really looking forward to the Nano set. Keep them coming Paul😊👍
Great lesson as always Paul !! First i couldn't get the piCam to work (after an hour i found that i had put the ribbon in the wrong way), then it worked ,although the picture is quit grainy......but hey it works. then i dug up an old microsoft vx-7000 webcam just for kicks...and ....it works too on the nano ;)
Hello apple thanks for letting us know how you fixed your picam apparently i had my ribbon cable flipped as well. I was just about to give up on my picam and looked at my ribbon and it was also flipped. Thank for the heads up.
This seemed the most appropriate place to ask about Logitech Webcam issues. I purchased the exact same camera as per your link, and I'm having issues with lagging on my jetson nano. Even with nothing but the basic camera program running. Any ideas?
How would you go about steaming live video over the network. Like rtsp or udp. When I get more advanced I would like to build a bot or rover with a first person view on another device. Or have the nano in another room and view the video on a phone, tablet, and or pc.
Hey Paul , you are my most excellent programming teacher. I had a question that I wish to use two raspberry pi cameras on my Jetson nano b01 model . What code should I write to run them simultaneously?
I’m trying this with opencast 4.1.1 and it’s throwing an error for the imshow line, from what I can tell it’s saying it can’t find the camera. Any suggestions? Thanks for all you do.
for IMX219 users: dispW = 816, dispH=616 and framerate 15/1 helped me to run the camera. I'm using Jetson Nano 2GB. 1200x900 works fine as well. I think maybe it just doesn't want to display in resolution that is too low or that is much smaller than capture resolution or something
Thank you so much for the tutorials. I am using the C920 Webcam and it appeared to be really lagging and the window sized could not be changed. I noticed that the version of OpenCV I have is 4.1.1 not 3.3.1, which is the problem. This happens to later user of Jetson Nano. I hope you find this comment and made some adjustments to your tutorials.
I love your videos. Very helpful. I am not new to Linux (NOT an expert) and I have quite a bit of development experience, so some of the content for me is rather mundane, but you did a great job explaining this stuff to someone VERY new. I am running a Jetson Orin Nano Dev (Yes, I am new to this platform), and so far everything you had us do works here. My question is: I have two IMX219 cameras. One is 77 fov and the other is 160 fov. How do I get BOTH cameras to show up?
Hi, Excellent tutorial! I have a question, for the new model that has 2 CSI ports, how must be the string of camSet for the second one? thanks for the support!
Look at my lessons on the Jetsan Xavier. It has two camera ports, and I show how to address the different ports in those lessons. I am thinking that should be the same with the new nanos.
Hi @@paulmcwhorter, I find that in 'camSet' you must indicate the index (Ex: sensor_id=0 or 1) but don't works with the version 4.1.1. I purge this version of OpenCV and install the version 3.1.1 and BOOM!, works perfectly. Your content is top and your videos are very well explicated. Thanks!
I know someone asked this question a year ago but didn't get any reply. I was wondering if it is possible to hook up a PoE/IP camera on jetson nano and use it for object recognition? I have a couple of ANPVIZ IPC-D250G-S IP cameras. Any help or guidance will be much appreciated. Thanks for the videos, Paul. These are absolutely priceless!
Hey Paul. great video, I really Appreciatte if you prepare a video in the future using the jetson nano with the opencv but runing on ubuntu without the GNOME GUI to optimaze memory and processor resources, and accessing the jetson from the network. Thank you very much indeed, great channel.
Everything went good until runnig the script for the camera ... It says assertion failed in 'imshow' and I can't find a good solution. No solution posted online on this error seemed to work. Help.
I just got my gear set up yesterday January 23, 2020. Using a Logitech C920S camera (the C920 has apparently been discontinued). The current image downloaded from the nvidia site has openCV version 4.1.1. The first time I ran the video 11 "base" program, the frame displayed full screen ??? After quitting the program, I saw a GStreamer warning message: Cannot query video position: status=0, value=-1, duration=-1. Huh??? Running Cheese produces a nice, expected frame. Did a little research and found that if I explicitly set the frame width and height before the while loop as follows: cam=cv2.VideoCapture(0) # existing line in the supplied base source cam.set(cv2.CAP_PROP_FRAME_WIDTH, dispW) # new line; uses dispW cam.set(cv2.CAP_PROP_FRAME_HEIGHT, dispH) # new line; uses dispH while True: # loop is unchanged then the frame displays properly. I don't know why I have to do this, but I'm passing the information along in case someone else runs into the problem. I plan to continue researching why I got the GStreamer warning message, but am not hopeful.
I finally got this lesson to work with cv2. version 4.1.1. What I did was to install "sudo apt-get install python3-matplotlib. Then commented out the print(cv2.__version__),in the openCV1.py script. Then followed the steps in the lesson, and it worked,(successful). Thanks Paul, Great job!
Good Morning Paul. Another great video, but I do have a question. I have 2 rPi cameras and would like to use both here on the Nano. WI don't have a second ribbon connector so ..Is there an adapter cable I can get to plug it into a USB slot? (I assume that is where you plugged in the web cam.) Thanks again for the great work. I would never have been able to get this far without these lessons.
What are you doing sir!!! Awesome . I am really impressed for your vast knowledge..I will definitely contribute to your petreon a/c in future if possible..thank you very much..carry on your best work..
Thanks for your tutorial! I now know what I'm doing when using functions in opencv. Btw, my picam is dead so I'm using the cam from my android phone via lan.
Thank you for this excellent guide. It has helped me get the camera working on the Nano but the quality is not very good. There is a pink tinge to the image. There are various postings on the web which suggest that other people have this problem, but I cannot see any way of linking the fairly baffling solutions to the setup which you have given.
Update to my previous post about problems with Logitech C920S camera when using the 2019/12/17 SD card image. Since my original post, I ran into another problem -- no "code completion" when typing cv2 related statements. For an example, look in the video at 10:42, where "cam=cv2.Video" is typed; note the popup that appears showing suggestions for "code completion". On my nano, I do NOT get code completion when using the 2019/12/17 image. This problem, coupled with the Logitech C920S problem described in a separate post, motivated me to post questions to the nvidia help forum. Links to these follow. Code completion post: 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/ Webcam post: devtalk.nvidia.com/default/topic/1071003/jetson-nano/opencv-4-1-1-errors-using-webcam/ At the same time, I decided to test using the 2019/07/19 SD card image, which is the one Paul is using. When I did so, both problems disappeared. I now have "code completion", no warnings using the webcam, and good initial values for frame width, height, and frame rate. Based on these results (and lack of response on the nvidia forum), I have switched permanently to the 2019/07/19 image. I would be interested in hearing anyone else's experience with the 2019/12/17 image. Did you have any problems?
I'm up and running , I have the new Jetson Nano with two camera. i can only make one work at any point in time. How do i get two Pi camera working. and where can i buy your book
Hello, i couldn't buy pi-camera or Logitech C920 WEB cam. I instead bought Logitech 4k brio. unfortunately, i used the code you wrote, but the frameratre is slow and the image depicted is moving slow. do you have any suggestion of how i can improve the video captured?
cam = cv2.VideoCapture('/dev/video1') instead of cam = cv2.VideoCapture(1) For window dimension install: pip3 install imutils # and add in your python program on first line: import imutils # and after ret, frame = cam.read() line insert another line with: frame=imutils.resize(frame,width=480) # you can change the window dimension from 480 to whatever.
another great content, althought from hardware point of view more challenging. One question, I have 2 piCams and want to test both at the same time. I have found some insight on the web i.e. adding sensor-id to the pipeline. but it does not work. Any suggestion?
I got it working, and the video is displayed correctly only vertically...its still flipped horizontally. I looked at all the codes 1-8 but I can't find the right orientation. Its a mirror image..I move my hand goin right and the camera shows it going left. Any tips where to start looking? Developer forums are just talking about the orientation issue that you have already resolved here (upside down), but nothing on the horizontal flip. Thank you for the great series!
Hello Mr. Paul McWhorter I have the imx219-77 camera connected to my jetson nano. Camera works fine from command line (using the command: nvgstcapture) but i can't open the camera with openCV. Any help? Thanks a lot for your great video tutorial! :-)
I am new to AI but I know programming. I have one question. I am planning to develop image recolonization (people) software where data is coming from IP cameras. Is it possible? Also, is jetson the best platform for that? Thanks Paul. you are the best.
I am using cv2 version 4.1.1 for Python3. I am receiving this warning which is a known cv2 issue: "GStreamer warning: Cannot query video position: status=0, value=-1, duration=-1".
Kamal, the error does not cause a problem. Things should work as expected. My videos were for jetpack 4.2 and it sounds like you are on jetpack 4.3. The only difference will be that in openCV when you do the findContours command, it only returns two parameters, so you have to adjuat your findContours commands for as follows _, contours, _ = should become just contours, _ =
@@paulmcwhorter Hi Mr. McWhorter, first thanks a lot for these interesting video tutorials, I really enjoy learning with you BOOM yes, the warning message does not cause a problem, I used the jetpack 4.3 and I installed OpenCV by compiling it from source (this enables me to use the GPU(cuda functions)) this is the complete waning message [ WARN:0] global /home/kamal/opencv/modules/videoio/src/cap_gstreamer.cpp (933) open OpenCV | GStreamer warning: Cannot query video position: status=0, value=-1, duration=-1 I will search for the warning message if I find a solution I will share it with you Thanks again
You are probably not reading the image from the camera because I think that parameters for that command changed in later versions of opencv. If you use the same version of openCV that I do, things should work.
Hi Paul, Let me start by complimenting you on these outstanding lessons. They are fantastic, and I think that you're doing a great job. I recently purchased my Jetson Nano (v.945-13450-0000-100) and installed this image (jetson-nano-jp451-sd-card-image.zip) onto it. My Python version is 3.6.9, and my OpenCV version is 4.1.1. I am wondering if this tutorial and the ones that follow will work with my installed versions of Python and OpenCV? If not do you have a tutorial explaining how to downgrade my versions? Thanks
hmm you're using a nano 2Gb in your videos , yet your link goes to a nano 4GB ,I can't get the camera to work I'm wondering if that's the problem ? does it have to be specified which camera port on the 4GB nano is being used as it has 2 camera ports on it not 1 like on the nano you're using
Hi all, Im using openCV ver 4.1.1 and this code doesnt work for me, is any one having the same problem? someone knows how to wake the camera in an other way? i tried few other videos here in youtube and none of them works. by the way, the camera is working correctly, i tried it on 'ras pi4' and it works.
Hi Paul, I've been working my wat through the Jetson Nano AI Tutorials to try and catch up for when you do the JetBot in the future. My Jetson Nano is very new and when I went through all the setup procedures it came with OpenCV 4.1.1, not the OpenCV 3 you have during this. My camera is returning two black and white pictures diagonally separated in the same display window. Do you have any idea how to fix this? It must be something that has changed in Gstreamer settings!? Thanks for the tutorials they are great. Best wishes Paul
OpenCV version you have is OK. if you are using a WEB cam, make sure you launch it with camera settings '/dev/video1' not just a '1'. Make sure your camera is compatible with the Jetson Nano.
@@paulmcwhorter Thanks for replying. I have the Raspberry Pi camera. the error I get is this: [WARN:0] global /home/nvidia/host/build_opencv/modules/videoio/src/cap_gstreamer.cpp (933) open OpenCV | GStreamer warning: Cannot query video position: status=0. value=-1, duration=-1. I've trying to find out what is causing the error on forums but not much luck yet. Does this mean anything to you? Thanks again! Paul
After getting a couple of error messages and finding and fixing them, my camera is working!!! On a side note to everyone, who does this after me, in order to make the q command work, make sure you click on the piCam image before typing q. If you do this it will close the camera window.
Hi, I've run into a problem here. when I import, I have a name error: name 'cv' is not defined This is what I've tried. 1. Adding "python.linting.pylintArgs": ["--extension-pkg-whitelist=cv2"] in the preferences 2. "python.linting.pylintArgs": ["--generate-members"] in the preferences 3. from cv2 import cv2 4. re installing & removing opencv I assume this is something with pylint, but cannot figure out what the problem is. Please help!
some say that the directory where openCV is installed is not correct, or the openCV is not installed. Sorry for asking because you will have hundreds of questions from all different peoples. =( I am thinking of restarting from JetPack installation and start all over...
Paul, Thanks a lot for your great lessons! In this lesson I did not get how your web cam was connected. I followed your lesson and immediately connected my Xiaomi smart home camera to jetson nano via usb. I tried VideoCapture(0-5). It did not work. Could you please explain this moment?
am following your videos and i found it helpful.but i kept getting this error when i run same command am using same device but my camera is camera-night vision CAM001
Keeps hanging up on line 11 says incorrect syntax I copied it as per but it doesn't work but at least the string isn't throwing up an error anymore but it still won't turn on the camera .I give up
Got everything in lesson 11 working except the 'sudo apt-get install libcanberra-gtx-module' command would return a 'E: Unable to locate package libcanberra-gtx-module'. I was all over the web trying to fix this and gave up since it was not a critical step.
Hi Paul your fix for the libcanberra error does not work my nano is coming back saying I should try to update . I have not done this as you say to not do any updates.
you might try sudo apt-get update. Then try the command again. if the error does not go away, it is not important, things still work. I just dont like errors.
Most things should be the same. When you get to the lesson on 'contours' the contour command has to be adjusted a little for your version, but think that is explained in the comments section
I've been through these lessons up to 28 a couple times now on my laptop. I rewrote this set every time just to make sure I got it down. I can write the whole thing on my own now and understand what each line does.
Just got my jetson nano gear and moving forward from here on and finally moving past lesson 28.
Great job!
That cup of coffee seems like heaven. Really appreciate the playlist. Thank you.
It was!
Who else is donating to Paul and his magnificent, freely available Arduino / Jetson / tech hobby video tutorials!? Everyone!!! Good! Three bucks a month is all it takes, five, eight, ten, you're getting a bargain in online education with Paul and his material.
I think I learn more from your errors than when you enter the code correctly. Thanks!
You know, that's a great point.
Everything up running well. Was surprised as was thinking a lot has changed since you posted the video. cv2 version ended up being 3.2.0 but works here for Lesson 11. I do have the Nano with two camera connections. Going to try and use the Luxonis Oak-D Camera with Nano. Thanks for all you do Paul. Great stuff.
Got the camera working great, looking forward to doing more stuff, i would never have been able to do this without your lessons
Not sure if this video is still monitored; but been enjoying the lessons and was following along so well that I actually paused the lessons until my Raspberry Pi camera came - even though I also had a compatible Logitech camera. Thank you for the lesson plan.
I love so much the fact that you use 둘글레차 as kind of focus meter😂
Thank you very much for this tutorial, had some challenges with getting openCV to work, but upgraded per Lesson 36 and I'm back on track. Boom!! :-)
The camera was exciting , I love it when you hold your breadth , you look superb. Your videos are super col and they explain more than just copying and pasting. I love em' all
Thank you so much sir.
These tutorials are incredible.
Your future supporter from Iran.
i got everything to work, and i am very exited about the possibility's with jetson. keep up the good work man, your an excellent teacher.
-1= == 30 mins looking for what wasn't there.
Thought I was going crazy, but I did learn to read error messages better!
Camera up & runnning!!!
"But enough of this shi...shameless promotion..." 😂Lol. Seriously though, great lesson Paul👍
😂😂😂
Lol
I don't think Paul is capable of cursing!
🤣🤣
omg loooool xD
Thank you again for such a groundbreaking lesson.
Working good so far. Very happy and exciting about the rest of lectures. Thank you, Paul.
Enjoying all of your course, Arduino, Python, now Nano. However, with the nano, sudo apt-get install python3-opencv the install is of 4.4.1 and it does not cooperate. Any suggestions or direction would be great as I look forward to completing this series. Is there any method to follow for an install of 3.3.1? Thanks
Newegg had $20 off the Nano so I ordered one so should have one soon to follow your series. Keep up the great work.Your excitement is contagous!
I really enjoyed this lesson...I'm excited to do more with openCV. Ordered my Py cam. Thanks Again!
Great lesson Paul. Always enjoy live experimentation...lol. Also cool to see 2 cameras in action. Many thx.
Hi Paul. I just wanted to say that I've watched all 11 lessons thus far and I am really excited about this course. For reasons I won't go into I haven't been able to follow you on the Jetson Nano, so I'll be playing catch-up this week. Hopefully I'll be all caught up by lesson 12. I have the hardware all set up and running fine. I'm also already familiar with Linux, Python, and Numpy so those will be easy for me. This will be my first experience with OpenCV.
At the end of this lesson you mentioned being able to recognize a specific person's face. This is actually my ultimate goal with this project. So I'm really looking forward to learning how to achieve that goal.
Thanks for creating these wonderful lessons and I'll be signing up on your Patreon site very soon. You definitely deserve to be compensated for your work. Superb instructions to be sure! Thank you so very much!
I'm all caught up. BOOM! Everything works great. Thank you for creating this superb series Paul. I'll be coming over to patreon shortly to support your hard work.
Things will start getting more interesting. I know it is taking a long time to get to the good stuff, but really excited about what this board can do.
@@paulmcwhorter That's cool. I'm in no rush. I just wanted to get caught up. I'm working with other things that have to do with AI too. I've been working with numpy and artificial neural networks. So this is just adding hardware to experiment with on top of that. I'm building a robot and one of the things I would like to robot to do is to be able to recognize specific individuals. I see you mentioned that possibility here at the end of this video. So I'm really looking forward to learning how to do that. But like I say, no rush. I'm just happy to be caught up.
That is the direction we are going. First, we will recognize things based on cascades, or training that others have already done, so we can get the nano to do things that are cool by download training sets others have done. Those training sets will allow us to do things like recognize cats verses dogs verses bananas and so forth. Then finally, the final challenge will be for us to create training cascades ourself. This will allow us to recognize a specific person. That is a tall order, but where we are going. I think the tough part of getting an AI system going is the training. Anyway, that is what I am working on. Feel a certain amount of pressure as I have to stay ahead of you guys, so that keeps me on my toes.
@@paulmcwhorter "This will allow us to recognize a specific person. That is a tall order, but where we are going" That's cool. I was going to try to do this on my own. So anything you have to offer will be greatly valued. Just setting the system up this far would have been a nightmare without your help. And I'm sure I'll learn much more from your lessons as well.
"Feel a certain amount of pressure as I have to stay ahead of you guys, so that keeps me on my toes." No pressure here. I'm extremely patient. I'm in no rush. Just really happy that you chose to do this course. This course is EXACTLY what I was hoping to find on TH-cam, so I couldn't be more pleased with your choice of subject matter. I think (no scratch that), I KNOW that I couldn't find anything as well organized as what you have presented thus far. You've taken us from the bare-bones assembly right up to getting the camera running being sure to preserve the 128 core capability of the Nano. Can't beat that! I owe you big time, and I'll be over to Patreon very shortly to settle up my debt. I might even do that tonight. But if not tonight, very soon to be sure. You deserve to get paid for all the work you've done. And your product (lessons) are very well worth it. No question about it. In fact, I don't even see how you can do so much as I know you are doing other lessons too. I'm even following some of those as well. You're a hard worker Paul. No question about that.
hello Paul thanks for the lessons i am really enjoying this series of lessons. I have always wanted to play with AI. Keep up the awesome video lessons.
Another great lesson. I thought I would have problems - I'm using the later B01 board with two camera ports, my openCV is 4.4.1 and well Boom all up and working no probs. Sad to have just finished the Fusion 360 lesson set but really looking forward to the Nano set. Keep them coming Paul😊👍
I have problem with openCV 4.4.1 Version, did you need to change the code?
Great lesson as always Paul !!
First i couldn't get the piCam to work (after an hour i found that i had put the ribbon in the wrong way), then it worked ,although the picture is quit grainy......but hey it works.
then i dug up an old microsoft vx-7000 webcam just for kicks...and ....it works too on the nano ;)
Hello apple thanks for letting us know how you fixed your picam apparently i had my ribbon cable flipped as well. I was just about to give up on my picam and looked at my ribbon and it was also flipped. Thank for the heads up.
I have geric csi camera.jetson is not detecting it but usb cam is ok.do i need official pi camera or am i missing something
This seemed the most appropriate place to ask about Logitech Webcam issues. I purchased the exact same camera as per your link, and I'm having issues with lagging on my jetson nano. Even with nothing but the basic camera program running. Any ideas?
Never mind. I solved it with this written in right under cam=cv2.VideoCapture(1) I wrote in...
Cam.set(cv2.CAP_PROP_FPS, 30)
Thanks so much for posting this
I wanted to thank you because when i get stuck i look at your videos and i get unstuck. Thanks.ps. i bought a bumper sticker to help support you.
How would you go about steaming live video over the network. Like rtsp or udp. When I get more advanced I would like to build a bot or rover with a first person view on another device. Or have the nano in another room and view the video on a phone, tablet, and or pc.
Hey Paul , you are my most excellent programming teacher. I had a question that I wish to use two raspberry pi cameras on my Jetson nano b01 model . What code should I write to run them simultaneously?
Look at my lessons on the Jetson Xavier they have some video on running two pi cameras
Thanks alot Paul ❤️
I’m trying this with opencast 4.1.1 and it’s throwing an error for the imshow line, from what I can tell it’s saying it can’t find the camera. Any suggestions? Thanks for all you do.
I have the same issue did you find anything?
Same here!
for IMX219 users: dispW = 816, dispH=616 and framerate 15/1 helped me to run the camera. I'm using Jetson Nano 2GB. 1200x900 works fine as well. I think maybe it just doesn't want to display in resolution that is too low or that is much smaller than capture resolution or something
for me its running bro
Thank you so much for the tutorials. I am using the C920 Webcam and it appeared to be really lagging and the window sized could not be changed. I noticed that the version of OpenCV I have is 4.1.1 not 3.3.1, which is the problem. This happens to later user of Jetson Nano. I hope you find this comment and made some adjustments to your tutorials.
Did you find a solution to your problem?
ya i m having the same problem. Pi camera works fine but logitech c920x camera is supper laggy. any solution?
I love your videos. Very helpful. I am not new to Linux (NOT an expert) and I have quite a bit of development experience, so some of the content for me is rather mundane, but you did a great job explaining this stuff to someone VERY new. I am running a Jetson Orin Nano Dev (Yes, I am new to this platform), and so far everything you had us do works here. My question is: I have two IMX219 cameras. One is 77 fov and the other is 160 fov. How do I get BOTH cameras to show up?
I am really exited to learn OpenCV by your way of teaching sir🤩🤩🤩🤩🤩
Keep watching
Hi,
Excellent tutorial! I have a question, for the new model that has 2 CSI ports, how must be the string of camSet for the second one? thanks for the support!
Look at my lessons on the Jetsan Xavier. It has two camera ports, and I show how to address the different ports in those lessons. I am thinking that should be the same with the new nanos.
Hi @@paulmcwhorter,
I find that in 'camSet' you must indicate the index (Ex: sensor_id=0 or 1) but don't works with the version 4.1.1. I purge this version of OpenCV and install the version 3.1.1 and BOOM!, works perfectly.
Your content is top and your videos are very well explicated. Thanks!
I know someone asked this question a year ago but didn't get any reply. I was wondering if it is possible to hook up a PoE/IP camera on jetson nano and use it for object recognition? I have a couple of ANPVIZ IPC-D250G-S IP cameras. Any help or guidance will be much appreciated. Thanks for the videos, Paul. These are absolutely priceless!
Did you ever figure out how to use other cameras?
Hey Paul. great video, I really Appreciatte if you prepare a video in the future using the jetson nano with the opencv but runing on ubuntu without the GNOME GUI to optimaze memory and processor resources, and accessing the jetson from the network. Thank you very much indeed, great channel.
Great session - LUV IT - Both cameras going . . .
Everything went good until runnig the script for the camera ... It says assertion failed in 'imshow' and I can't find a good solution. No solution posted online on this error seemed to work. Help.
well, i found the problem. If I ran the script in python 3.6 it displayed the error, but when I run it in 2.7 it works. Still don't know why
I just got my gear set up yesterday January 23, 2020. Using a Logitech C920S camera (the C920 has apparently been discontinued). The current image downloaded from the nvidia site has openCV version 4.1.1. The first time I ran the video 11 "base" program, the frame displayed full screen ??? After quitting the program, I saw a GStreamer warning message: Cannot query video position: status=0, value=-1, duration=-1. Huh??? Running Cheese produces a nice, expected frame. Did a little research and found that if I explicitly set the frame width and height before the while loop as follows:
cam=cv2.VideoCapture(0) # existing line in the supplied base source
cam.set(cv2.CAP_PROP_FRAME_WIDTH, dispW) # new line; uses dispW
cam.set(cv2.CAP_PROP_FRAME_HEIGHT, dispH) # new line; uses dispH
while True:
# loop is unchanged
then the frame displays properly. I don't know why I have to do this, but I'm passing the information along in case someone else runs into the problem. I plan to continue researching why I got the GStreamer warning message, but am not hopeful.
I see this problem as well with a RPi Cam V2. Adding the code doesn't work for CSI cams I think.
This worked for me as well. Using Logitech c920s pro
John would you mind posting your CAM string to build the pipeline? I'm trying to get the C920 working as well.
I finally got this lesson to work with cv2. version 4.1.1. What I did was to install "sudo apt-get install python3-matplotlib. Then commented out the print(cv2.__version__),in the openCV1.py script. Then followed the steps in the lesson, and it worked,(successful). Thanks Paul, Great job!
Perfect! Solved my problem
not works for me :(
19:20 둥글레차 ㅎㅎ
if anyone is facing issues with the camera working. Make sure that the camera is attached to the jetson before turning it on
Good Morning Paul. Another great video, but I do have a question. I have 2 rPi cameras and would like to use both here on the Nano. WI don't have a second ribbon connector so ..Is there an adapter cable I can get to plug it into a USB slot? (I assume that is where you plugged in the web cam.) Thanks again for the great work. I would never have been able to get this far without these lessons.
I dont know any way to do that.
@@paulmcwhorter The Jetson Nano B01 version has two camera connectors
Does all the code covered by you in this course be used in PC too, instead of the Jetson Nano?
What are you doing sir!!! Awesome . I am really impressed for your vast knowledge..I will definitely contribute to your petreon a/c in future if possible..thank you very much..carry on your best work..
Thanks for your tutorial! I now know what I'm doing when using functions in opencv. Btw, my picam is dead so I'm using the cam from my android phone via lan.
Thank you for this excellent guide. It has helped me get the camera working on the Nano but the quality is not very good. There is a pink tinge to the image. There are various postings on the web which suggest that other people have this problem, but I cannot see any way of linking the fairly baffling solutions to the setup which you have given.
how are u sir.. i want to ask you.. i want to install pyrealsense2 on my jetson nano.. is there anyway to do it??
Update to my previous post about problems with Logitech C920S camera when using the 2019/12/17 SD card image. Since my original post, I ran into another problem -- no "code completion" when typing cv2 related statements. For an example, look in the video at 10:42, where "cam=cv2.Video" is typed; note the popup that appears showing suggestions for "code completion". On my nano, I do NOT get code completion when using the 2019/12/17 image.
This problem, coupled with the Logitech C920S problem described in a separate post, motivated me to post questions to the nvidia help forum. Links to these follow.
Code completion post:
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/
Webcam post:
devtalk.nvidia.com/default/topic/1071003/jetson-nano/opencv-4-1-1-errors-using-webcam/
At the same time, I decided to test using the 2019/07/19 SD card image, which is the one Paul is using. When I did so, both problems disappeared. I now have "code completion", no warnings using the webcam, and good initial values for frame width, height, and frame rate. Based on these results (and lack of response on the nvidia forum), I have switched permanently to the 2019/07/19 image.
I would be interested in hearing anyone else's experience with the 2019/12/17 image. Did you have any problems?
I'm faced with the same issue. Like you mentioned, i might try using an older image file
Very cool stuff, so excited about the upcoming videos 😎🤩🤩🤩🤩
I'm up and running , I have the new Jetson Nano with two camera. i can only make one work at any point in time. How do i get two Pi camera working. and where can i buy your book
How do you change camera port or add another camera
Hello, i couldn't buy pi-camera or Logitech C920 WEB cam. I instead bought Logitech 4k brio. unfortunately, i used the code you wrote, but the frameratre is slow and the image depicted is moving slow. do you have any suggestion of how i can improve the video captured?
make sure to launch camera with 'dev/video1' not just a '1'. Then set the camera resolution lower if needed using opencv camera set command.
cam = cv2.VideoCapture('/dev/video1') instead of cam = cv2.VideoCapture(1)
For window dimension install:
pip3 install imutils
# and add in your python program on first line:
import imutils
# and after ret, frame = cam.read() line insert another line with:
frame=imutils.resize(frame,width=480)
# you can change the window dimension from 480 to whatever.
@@paulmcwhorter Sometimes 'dev/video1' may not work so just replace the 1 with a 0 so '/dev/video0'
Very interesting! Love it! Great tutorial!
Thanks so much!
Nvidia should be sponsoring you.
how do we fix the lag on the other camera?
another great content, althought from hardware point of view more challenging. One question, I have 2 piCams and want to test both at the same time. I have found some insight on the web i.e. adding sensor-id to the pipeline. but it does not work. Any suggestion?
Look at my NVIDIA Xavier playlist. I show the commands to run two piCams at once.
I got it working, and the video is displayed correctly only vertically...its still flipped horizontally. I looked at all the codes 1-8 but I can't find the right orientation. Its a mirror image..I move my hand goin right and the camera shows it going left. Any tips where to start looking? Developer forums are just talking about the orientation issue that you have already resolved here (upside down), but nothing on the horizontal flip. Thank you for the great series!
Hello Mr. Paul McWhorter
I have the imx219-77 camera connected to my jetson nano.
Camera works fine from command line (using the command: nvgstcapture) but i can't open the camera with openCV.
Any help?
Thanks a lot for your great video tutorial! :-)
This was a lot more troublesome trying to do headless.
Awesome tutorial!! But how do run the same thing on GPU? I mean to capture an image using CUDA in OpenCV-Python?
hi i want to take a picture with jetson nano python code how do i do?
So do you have information on changing the default web camera sizes? When I use mine it takes up the entire screen and makes it so slow
mine does the same thing did you ever figure it out?
I am new to AI but I know programming. I have one question. I am planning to develop image recolonization (people) software where data is coming from IP cameras. Is it possible? Also, is jetson the best platform for that? Thanks Paul. you are the best.
I think there is a mistake in the native resolution of the raspberry v2 camera, I believe it should be 3280 x 2464.
I am using cv2 version 4.1.1 for Python3. I am receiving this warning which is a known cv2 issue: "GStreamer warning: Cannot query video position: status=0, value=-1, duration=-1".
Me too with the same configuration I am getting the same error, have you find solution for that problem?
Kamal, the error does not cause a problem. Things should work as expected. My videos were for jetpack 4.2 and it sounds like you are on jetpack 4.3. The only difference will be that in openCV when you do the findContours command, it only returns two parameters, so you have to adjuat your findContours commands for as follows _, contours, _ = should become just contours, _ =
@@paulmcwhorter
Hi Mr. McWhorter, first thanks a lot for these interesting video tutorials, I really enjoy learning with you BOOM
yes, the warning message does not cause a problem, I used the jetpack 4.3 and I installed OpenCV by compiling it from source (this enables me to use the GPU(cuda functions))
this is the complete waning message
[ WARN:0] global /home/kamal/opencv/modules/videoio/src/cap_gstreamer.cpp (933) open OpenCV | GStreamer warning: Cannot query video position: status=0, value=-1, duration=-1
I will search for the warning message if I find a solution I will share it with you
Thanks again
In 2022 not works for me. I'm using cv2 4.1.1 and python 3.6.9. I have error in function "imshow"
You are probably not reading the image from the camera because I think that parameters for that command changed in later versions of opencv. If you use the same version of openCV that I do, things should work.
Hello, I am working on cv2 4.1.1 version. Did you solve this error on 4.1.1 version?
Hi very good video thank you. I am using JupyterLab as IDE but when I run the code, camera is not displayed, why is this?
I tried a lot of things but cant resolve this problem, when i tried to run it says "no camera aviable", how can i fix it?
thanks for clear explanation. BOOM!
Hi sir,how to take image with raspary pi web cam and save it on jetson nano????
Hi Paul,
Let me start by complimenting you on these outstanding lessons. They are fantastic, and I think that you're doing a great job.
I recently purchased my Jetson Nano (v.945-13450-0000-100) and installed this image (jetson-nano-jp451-sd-card-image.zip) onto it.
My Python version is 3.6.9, and my OpenCV version is 4.1.1.
I am wondering if this tutorial and the ones that follow will work with my installed versions of Python and OpenCV?
If not do you have a tutorial explaining how to downgrade my versions?
Thanks
hmm you're using a nano 2Gb in your videos , yet your link goes to a nano 4GB ,I can't get the camera to work I'm wondering if that's the problem ? does it have to be specified which camera port on the 4GB nano is being used as it has 2 camera ports on it not 1 like on the nano you're using
I am using a 4GB nano, but an earlier board with one camera port. You need to specify which camera if you have two ports on the newer boards.
Hi all,
Im using openCV ver 4.1.1 and this code doesnt work for me, is any one having the same problem? someone knows how to wake the camera in an other way? i tried few other videos here in youtube and none of them works.
by the way, the camera is working correctly, i tried it on 'ras pi4' and it works.
Yeah, the code is not working with openCV ver 4.1.1. I am still looking for solution to sort it out.
Hi Paul, I've been working my wat through the Jetson Nano AI Tutorials to try and catch up for when you do the JetBot in the future. My Jetson Nano is very new and when I went through all the setup procedures it came with OpenCV 4.1.1, not the OpenCV 3 you have during this. My camera is returning two black and white pictures diagonally separated in the same display window. Do you have any idea how to fix this? It must be something that has changed in Gstreamer settings!? Thanks for the tutorials they are great. Best wishes Paul
OpenCV version you have is OK. if you are using a WEB cam, make sure you launch it with camera settings '/dev/video1' not just a '1'. Make sure your camera is compatible with the Jetson Nano.
@@paulmcwhorter Thanks for replying. I have the Raspberry Pi camera. the error I get is this: [WARN:0] global /home/nvidia/host/build_opencv/modules/videoio/src/cap_gstreamer.cpp (933) open OpenCV | GStreamer warning: Cannot query video position: status=0. value=-1, duration=-1. I've trying to find out what is causing the error on forums but not much luck yet. Does this mean anything to you? Thanks again! Paul
Sir, thanks but how did you connect webcam?
Go to video #1 in this series.
I like your lessons. I have been using Arduinos. I'm not very good at Python programing. I use thought programming per your examples.
My tablet is junkie I found it thanks Paul.
After getting a couple of error messages and finding and fixing them, my camera is working!!! On a side note to everyone, who does this after me, in order to make the q command work, make sure you click on the piCam image before typing q. If you do this it will close the camera window.
Thanks!!! That worked for me.
Hi, I've run into a problem here.
when I import, I have a name error: name 'cv' is not defined This is what I've tried.
1. Adding "python.linting.pylintArgs": ["--extension-pkg-whitelist=cv2"] in the preferences
2. "python.linting.pylintArgs": ["--generate-members"] in the preferences
3. from cv2 import cv2
4. re installing & removing opencv
I assume this is something with pylint, but cannot figure out what the problem is.
Please help!
my open cv version is 4.1.1
some say that the directory where openCV is installed is not correct, or the openCV is not installed.
Sorry for asking because you will have hundreds of questions from all different peoples. =(
I am thinking of restarting from JetPack installation and start all over...
I started all over from the beginning and worked just fine :) thankyou!
Paul, Thanks a lot for your great lessons! In this lesson I did not get how your web cam was connected. I followed your lesson and immediately connected my Xiaomi smart home camera to jetson nano via usb. I tried VideoCapture(0-5). It did not work. Could you please explain this moment?
Care to share if you found the solution to your problem.
thanks another great lesion .
am following your videos and i found it helpful.but i kept getting this error when i run same command
am using same device but my camera is camera-night vision CAM001
Keeps hanging up on line 11 says incorrect syntax I copied it as per but it doesn't work but at least the string isn't throwing up an error anymore but it still won't turn on the camera .I give up
Great lesson...BOOM!
Got everything in lesson 11 working except the 'sudo apt-get install libcanberra-gtx-module' command would return a 'E: Unable to locate package libcanberra-gtx-module'. I was all over the web trying to fix this and gave up since it was not a critical step.
The "gtx". Should be "gtk".
Hi Paul your fix for the libcanberra error does not work my nano is coming back saying I should try to update . I have not done this as you say to not do any updates.
you might try sudo apt-get update. Then try the command again.
if the error does not go away, it is not important, things still work. I just dont like errors.
had an error with imshow..... ended up my camera ribbon cable was backwards XD
This was great 👍🙂
Good evening Paul, how can I download OpenCV 3.3.1? last lesson it was install openCV 4.4.1 and I have an error with running camera
LEGEND!
@@paulmcwhorter what do you mean by LEGEND?
@@paulmcwhorter I figured, the problem was connection b/w nano and camera. it works fine on openCV 4.4.1
@@wongyuhan104Hello, I’m having an issue with getting the camera working. I have same version but getting errors in the code. Did you this problem?
I got an import error by opencv-python 4.5.1.48. I tried "from cv2 import cv2". Then it works.
AttributError: module cv2 has no attribute videoCapture, how can i solve this problem
Check your code, find your error
try "from cv2 import cv2"
2021 opencv version is 4.1.1 hopefully this lesson will work :)
Most things should be the same. When you get to the lesson on 'contours' the contour command has to be adjusted a little for your version, but think that is explained in the comments section
sorry, but all I can say now is that THE CAMERA WORKS! I DIDN'T BREAK IT WHEN INSTALLING IT! :)
I'm currently using a web cam c720 and my image is just a green screen. Can someone help, please?
Works Great!
Great!
just installed OpenCV 3.4 with the whole long process today and back to this lesson...the way in lesson 10 doesn't work for me.
If I understand right, the new disk image from nvidia has the 3.3.1 on it.
Great Content!
Hello, How can I get a snapshot and save it from raspberry pi camera?
use cheese app
see using python cv2.imwrite('nameofyourimage', yourimagematrix, imageparameters) for image parameters see imwriteflags in opencv