AI on the Jetson Nano LESSON 44: Running Two Cameras in OpenCV

[@normb2780]

Hi Paul. Love all your lessons. I don't have a webcam, so I used two pi cams since my version of the Jetson Nano has two pi connectors. I found it necessary to define two camSets: camSet1 and camSet2. In camSet1 I added "sensor_id=0 " (without the quotes) immediately after nvarguscamerasrc, and in camSet2 I addeded sensor_id=1 immediately after nvarguscamerasrc. Then cam1 was defined with camSet1 and cam2 was defined with camSet2. After that, I did everything as you instructed and the two pi cams worked fine (of course with the latency problem you discussed). It's great when things work, although I admit watching you "crash and burn" is loads of fun too. Thanks again for your great lessons.

[@prathiknarayan1093]

I struggled for many days as I was getting a very slow frame rate. After switching to LT4 and upgrading the dlib 19.21, I now have a fps of ~21 with both the picam and the webcam working in parallel. Gives me some motivation to proceed with the other lessons.

[@Bob-zg2zf]

26:51 wow, finally, Paul touches on Class. Super nice explanation for the concept of class.🥰

[@paulmeistrell1726]

Fun lesson Paul, Looking forward to new concepts. The 2g is definitely running slower than your machine. The pi cam is causing the slowing. Probably do to memory usage with 2g swapping a lot.

[@Mircea007]

Hi Paul, I'm using a cheap webcam and the pi cam and they are both in sync. Looking forward for the next lesson on functions and classes.

[@wayneswan3092]

So I tried to take these methods over to my other programs and combine my frames into a single window, first using "np.vstack" of 2 frames for a total of 4 frames in 2 windows. That all worked out fine, but then when I tried to combine those two "vstacks" into a single "hstack" for all 4 frames in one window, I got all kinds of errors, which I can barely understand to have to do with frame size. Doing a "resizeWindow" to make them all the same size didn't clear up the error. I'll try and screenshot my methods and the errors and post them to KZbin so you can see what I'm talking about.

[@AlanHeigl]

great show again!. can't wait to learn classes next week

[@geeksatlarge]

Nice! Can't wait to see the FPS is like once I have a "good" PSU. Currently running the typical 5V 2A, and started getting "System throttled due to over current" when both cams kick on with facial recognition going.

[@opalprestonshirley1700]

Good lesson. I would like to see how to intergrate the cameras with the robotics car. How would you like the robot to lock on to you and follow you around annoying you to no end...LOL. Have a great weekend Paul.

[@keithemerson9349]

Really looking forward to the python lessons. I've avoided classes up to know (mainly because of the _self_ stuff !). Also it would be great how to use threading properly. great lessons!

[@sanfinity_]

Great video sir. Excited to see the next one

[@paulmcwhorter]

Coming soon!

[@SarkarMotion]

Somebody gives you a more powerful computer, you immediately ramp up task load by X2! Holy moly!

[@Bob-zg2zf]

Combining your gyro project with this AI project to stabilize the camera would make this project perfect.

[@marksholcomb]

Excellent! Thank you.

[@fun_engineering5710]

Hi Sir, I just watched your live shop talk videos of banggood and math challenge. and I wondered if you can make videos about calculus, engineering 1, engineering 2, engineering 3, engineering 4, but with more focus on the real life application in the engineering field. I think with this we can do better projects ,since we can build them with numbers that are calculated to help make the project most efficient and productive. Thank you so much Sir.

[@kekocan2900]

Look at Khan Academy. They will provide you excellent math teaching.

[@SKTWoodDesign]

Can you make a cctv multicamera app with intrusiin detection

[@OZtwo]

I found if you replace the webcam with the full name (/dev/video1) it will keep the same size.

[@SarkarMotion]

What if the second camera is IR based? Can we do depth sensing with that kind of a setup? Can we set the color profiles of two normal cameras to different color_spaces in order to generate 3D depth?

[@paulmcwhorter]

Probably very doalbe

[@theteenengineer7589]

Hi Mr.Paul, the jetson nano link is not working, is there another link to buy a good Jetson nano kit? Thank you

[@paulmcwhorter]

This is what you want: amzn.to/2XHDeCe

[@donwagner8126]

Thanks!

[@paulmcwhorter]

Really appreciate that!

[@paulmeistrell1726]

Interesting the pi cam is faster on this by about 3 times, about 21-22 fps, web cam is 7.6 fps!

[@wishicouldarduino8880]

For all the power these boards have I haven't seen any ai projects made with them I'm not knocking them .but I haven't seen any robot using it a full robot walking talking thinking doing expressing opinion or otherwise when someone says ai that is what everyone thinks I just haven't seen any huge ai project using either one .my pointe is this for all of the ai robots out there the real ones what are they using ? Jetson is relatively new most of the big ai bots were there before it .so I wonder what they use since there is not a Jetson in a single one .just curious.👍 cool video.

[@keithemerson9349]

levelled out at 30 fps with a single webcam. Does that sound about right?

[@ushachandramohanan1187]

Sir can you do a new tutorials regarding raspberry pi

[@CodingScientist]

Hi Paul I am getting only 5FPS in NX, you were getting around 16 FPS, how ?

[@CodingScientist]

may be because ur using SSD ?

[@paulmcwhorter]

Something is wrong. What fps do you get using the old simple program with just one camera?

[@CodingScientist]

@@paulmcwhorter ahh I did not try that, let me check

[@tralfazy]

Anbu, I did the procedure to boot from an external USB drive on my Jetson Nano (from JetsonHacks.com). The result was a bad hit to performance! I only get 3.0 fps when running Paul's program for a single camera when booting from the external drive. When I put the other SD card back in the fps jumps way up by almost 3 times to 8.4 fps! I was very disappointed at this result. The NX has the option for plugging in a solid state drive. Do you have that? Does it lower performance if you boot from it?

[@opalprestonshirley1700]

There is a lag time between the cameras.

[@paulmcwhorter]

Which is what we will deal with in next week's lesson

[@keithemerson9349]

running at about 7.5 fps compared to your 16 fps. Can't be the face recognition this time. runs at 24 fps with a single webcam

[@petersobotta3601]

Me too. Running a web cam and pi cam... The pi cam is two seconds behind

[@petersobotta3601]

My Pi cam runs at about 20 fps by itself and the Logitech c270 webcam runs at 7.5 fps.

[@coffegrinder406]

Are you orthodox? Did you ever think about becoming orthodox?

[@paulmcwhorter]

I am a follower of Jesus Christ. I shun denominational labels, as the Bible teaches there is One Faith, One Spirit and One Baptism. Hence I am a believer, follower and yes I worship with other believers, but I renounce denominational labels.

[@coffegrinder406]

@@paulmcwhorter ok that is fine i respect whatever you say i just want to tell you that there are really smart orthodox monks like monk Arsenije Jovanovic if you want to you can look him up and see. God bless you, you really can transfer knowlege, i dont know if you remember me im that guy from your live i was joking about being related to Nikola Tesla, also i made an arduino calculator that is so good where can i send you a picture i want to share it with someone who understands how hard it was everybody just says meh its fine.

[@SarkarMotion]

More people have been liquidated in the name of religion than women or property! Therefore I run from any and all varieties. I keep an open outlook and listen to all good teachings. But I must follow the Great spirit, the builder, the observer. We all come from the manifest of a single benevolent vibration. Our divine prerogative is to experience creation. Anything we do to uphold its positivity is rewarded by this benevolence, and everytime we lend ourselves to destruction, we reduce our own power to remain connected to the eternal vibration. But without this guiding vibration, we are truly lost.