Wow, I can’t thank you enough for making this video! It’s exactly what I was looking for-AI object detection on the Raspberry Pi 5-and you absolutely nailed it. The way you explained everything made it so easy to follow, and it feels like you really put your heart into this. I’m so grateful you took my request seriously and turned it into something so helpful and inspiring. You’re honestly the best-please keep up this amazing work! 🙌😊
Thanks for the kind words, I'm glad you found it helpful! I am glad for the suggestion as I ended up really enjoying this and want to play more with these capabilities.
I think doing car and pedestrian counting with the "segment anything". Great tutorial. I think that Super Nano is a game changer, but it seems to be sold old everywhere. Thanks for a great tutorial!
Thanks so much! Great fun! I am setting up my Pi right along with you. Thanks for the Jetson instructions. Will work on that next. Cool that you are building the mini case. I did order the Yahboom case for the Jetson. I'll let you know if it is any good when it comes in.
Absolutely pleasant introduction into what I was interested in. Nothing to say but joining all the positive comments! Maybe the screenrecorded text should be better readable.
Thanks very much! I appreciate the feedback on the text. I will try to make it more readable next time by at least zooming the browsers in more or something like that.
Great couple of videos bruv on the jetsun. With your studio set up and ease of talking on the camera I'd say that you've definitely found a niche. Keep rolling with it,. Saved me a lot of headaches with loading the ssd memory. Still to order a camera for this tutorial. I was wondering which camera that you used for the Jetson. Was it the Raspberry Pi AI one or a different one? Just went over your tutorial again and see that you used a usb webcam for the tutorial. SO question answered, For me the two reasons for getting this product. One was the video capabilities of the Orin Nano module, but my main reason was to play about with an localised ai environment, as I have a mac the options are limited and it proved cheaper to go with the orin nano than buy a full PC. Anyhow best of luck to you and look forward to more tutorials.
Thanks for the kind words. I think you will enjoy the jetson and yes as you saw I used a usb webcam as I didn't have luck running that script as is with the raspberry pi camera. Cheers!
@@Bijanbowen there is a built in 3-band EQ in OBS if you right click the audio source -> filters -> audio filters. you can also use third party VSTs in OBS if you want to get real picky. not intending to give you a hard time at all. i just think the rest of your video production is excellent, so the audio should match! keep it up either way.
@@Sash2240 Thanks very much, I know no hard feelings I am very open to feedback and this is helpful. i will play with the obs filters and see the difference. I should probably break out the studio monitors to edit instead of the ancient set i am using now hahah
You should check out the YOLO World model for object recognition. Kind of similar to the nano it can recognize in a bit more real time without having to be trained on everything.
Hello mate, the use case for security cameras is amazing, im looking for courses on this topic but are either obsolete or not with the latest jetson nano features, please, if you could create a tutorial i will be glad to pay for it Jetson nano plus IP security cameras with AI😊 😮😮
I don't know RE fps. The jetson takes 25w and the pi uses a 27w psu, though I don't know about actual draw at the socket as I don't have a measuring device.
The telephoto lens could be used to spot rogue drones 💀. Seriously though, I wonder if it could be used in conjunction with the regular camera in a security type of configuration. If the main camera detects something unusual, it turnes on the telephoto camera which of course would be mounted on a pan and tilt (gimble?) so that it can track the subject.
First thought of mine was drone related as well haha. Interesting thought about using it sequentially with a simpler camera and that could definitely be an implementation of it.
I have a really interesting thought, don't know if it's even possible or not. I have been looking for good car stereo for a while and to my disappointment even top ones from Sony, Kenwood and Pioneer don't have a complete system which has at least a 1080p display, Android Auto, WiFi, bluetooth and supports 4 camera inputs for a 360 degree systems like in modern cars. So I started to look for DIY options including a Raspberry Pi but the best OpenAuto OS for this usecase shut down recently. And considering the image processing limitations of the Raspberry Pi, I started wondering if any of this would be possible on a Jetson Nano. With its Ai prowess, it can even have blind spot monitoring and other safety features as well. Thanks for making all these videos!
This is a very interesting thought. I believe it would be possible at least with blind spot monitoring, and a car would have no problem powering the nano as well. Very interesting idea, I do have a relatively decent project car in the garage that I could attempt some of these things with, I just don't know if anyone would be interested haha.
@@Bijanbowen I appreciate you even thinking about the implementation of this absurd idea lol. People don't even think about going this route, is perhaps the reason no one has had talked about it yet. But when they get to know they can make the fastest, most powerful and feature packed car stereo, it will blow up.
@@ashyouknow7420 It is an awesome idea. I have been toying with how to go about using this to put blind spot monitoring in my 1988 Saab ever since reading your comment haha
Sorry to hear. I see the usb camera input functionality is still in beta, but maybe this can provide some assistance, good luck and hope it gets solved! github.com/hailo-ai/hailo-rpi5-examples/blob/main/doc/basic-pipelines.md#usb-camera-input
Thank you! To be honest, I have not used frigate so I can't say for sure. Looking at the github for it it seems rather powerful and something worth investigating.
which is better Nvidiai jetson nano orion or RPI 5 with AI extension , like for transformer based computer vision , segementation , classification of realtime vision , and which camera module is suitable for it
Jetson will be "better", pi is easier to use but not as powerful. For the Pi you can use any of the rapsberry pi cameras and for the jetson it becomes more complicated and you pick depending on what your use case works with USB or CSI camera.
Thank for being interested in turning the Jetson into a gaming computer I was wondering if you could do it I would be interested in purchasing it if was possible
Thanks! That is just a gaming comp style case, unfortunately getting the jetson to actually game seems to be a rather difficult task that may not be worth it in the end.
Very cool example. I like Vit implementation but the VLM can do better right? Like describing the image. Im thinking to extend Frigate functionality or even write frigate inspired pipeline detection but do some improvement after CNN detection and pass to VLM. With Jetson nano, i think this is possible. Example use case as security camera , VLM can describe parcel delivery guy etc. if the small VLM listed on jetson website can return json , then can do many things with this. 100 local
Thanks very much! Yes, VLM can absolutely do more in terms of descriptive power and a VLM with a json function call to perform an action if X is detected would be very very cool and replace a lot of stupid subscription service home security cameras, even if it was just to send a text notification of movement or something like that.
I have a homestead and thinking about setting up a Pi 5 with a camera running AI with Python. Then ask the AI if it sees an image of my dog chasing the chickens to then play audio noise of me saying NO or STOP!
Wow, I can’t thank you enough for making this video! It’s exactly what I was looking for-AI object detection on the Raspberry Pi 5-and you absolutely nailed it. The way you explained everything made it so easy to follow, and it feels like you really put your heart into this. I’m so grateful you took my request seriously and turned it into something so helpful and inspiring. You’re honestly the best-please keep up this amazing work! 🙌😊
Thanks for the kind words, I'm glad you found it helpful! I am glad for the suggestion as I ended up really enjoying this and want to play more with these capabilities.
I think doing car and pedestrian counting with the "segment anything". Great tutorial. I think that Super Nano is a game changer, but it seems to be sold old everywhere. Thanks for a great tutorial!
Thanks for the kind words! Hopefully it is back in stock soon!
Thanks so much! Great fun! I am setting up my Pi right along with you. Thanks for the Jetson instructions. Will work on that next. Cool that you are building the mini case. I did order the Yahboom case for the Jetson. I'll let you know if it is any good when it comes in.
I am looking forward to hearing your thoughts on the case once you get it.
Big Thanks for the video - I am also getting excited by exploring it.
No problem! You and me both haha
Great hands-on content on Jetson. Thanks and keep up the good work.
Thanks very much, will do for sure!
Next GREAT Video 🎉 thx so much for this quality and fun!! Learning so much with all of these vids together
Thanks for watching and I am glad you are finding the videos to be helpful!
Bro, love this. Thanks for the vid. Subbed!
I appreciate it!
Thank you, that was great information.
Glad it was helpful!
Thanks and happy new year !
Thank you and to you as well!
Cool insight into both; raspberry is meant to be cheap while as jetson is made for performance
Thank you, yes you are correct there!
Absolutely pleasant introduction into what I was interested in. Nothing to say but joining all the positive comments! Maybe the screenrecorded text should be better readable.
Thanks very much! I appreciate the feedback on the text. I will try to make it more readable next time by at least zooming the browsers in more or something like that.
Great couple of videos bruv on the jetsun. With your studio set up and ease of talking on the camera I'd say that you've definitely found a niche. Keep rolling with it,. Saved me a lot of headaches with loading the ssd memory. Still to order a camera for this tutorial. I was wondering which camera that you used for the Jetson. Was it the Raspberry Pi AI one or a different one? Just went over your tutorial again and see that you used a usb webcam for the tutorial. SO question answered, For me the two reasons for getting this product. One was the video capabilities of the Orin Nano module, but my main reason was to play about with an localised ai environment, as I have a mac the options are limited and it proved cheaper to go with the orin nano than buy a full PC. Anyhow best of luck to you and look forward to more tutorials.
Thanks for the kind words. I think you will enjoy the jetson and yes as you saw I used a usb webcam as I didn't have luck running that script as is with the raspberry pi camera. Cheers!
Thanks and happy new year
Thank you and you as well!
With that telephoto camera, maybe tracking birds, cats, bugs... cars and other vehicles. Maybe stars if sensitive anough.
I would like to point it at the sky at night and see if it produces any clear images. Thanks for the suggestions.
Great! thank you!
Thanks for the support!
cool vid and thanks for the info. you may benefit from a high-pass eq filter on your vocals it's a bit muffled at low volumes.
Good suggestion, I am only recording through obs right now, but if I decided to swap over to using logic pro it should be a snap to do.
@@Bijanbowen there is a built in 3-band EQ in OBS if you right click the audio source -> filters -> audio filters. you can also use third party VSTs in OBS if you want to get real picky. not intending to give you a hard time at all. i just think the rest of your video production is excellent, so the audio should match! keep it up either way.
@@Sash2240 Thanks very much, I know no hard feelings I am very open to feedback and this is helpful. i will play with the obs filters and see the difference. I should probably break out the studio monitors to edit instead of the ancient set i am using now hahah
Good stuff, thanks.
You're very welcome!
wunderbar
Thank You!
THANKS. iNCREDIBLE. HAVE A GOOD 2025
Thanks very much and you as well!
We've trashed a camera on the jetson before by backwards connecting it (on a carrier board), but guess that wasn't your issue.
That's very interesting to know. I am lucky that wasn't my issue and that camera still works!
You mentioned that you created a tutorial for the Jetson. Where can we find that documentation?
ruclips.net/video/-PjMC0gyH9s/видео.html
You should check out the YOLO World model for object recognition. Kind of similar to the nano it can recognize in a bit more real time without having to be trained on everything.
Thanks for this recommendation, it looks interesting and I will be looking more into these technologies.
Hello mate, the use case for security cameras is amazing, im looking for courses on this topic but are either obsolete or not with the latest jetson nano features, please, if you could create a tutorial i will be glad to pay for it
Jetson nano plus IP security cameras with AI😊 😮😮
what is fps and power consumption on both device
I don't know RE fps. The jetson takes 25w and the pi uses a 27w psu, though I don't know about actual draw at the socket as I don't have a measuring device.
The telephoto lens could be used to spot rogue drones 💀. Seriously though, I wonder if it could be used in conjunction with the regular camera in a security type of configuration. If the main camera detects something unusual, it turnes on the telephoto camera which of course would be mounted on a pan and tilt (gimble?) so that it can track the subject.
First thought of mine was drone related as well haha. Interesting thought about using it sequentially with a simpler camera and that could definitely be an implementation of it.
thank you algo, and you ser
Thanks very much!
That Betty is pretty cute!!!
Woof!
Te!ephoto use case. How about positioning a bird feeder in the field of view and have a model that identifies various species that frequent for food.
That sounds like a good use case for that lens.
I have a really interesting thought, don't know if it's even possible or not. I have been looking for good car stereo for a while and to my disappointment even top ones from Sony, Kenwood and Pioneer don't have a complete system which has at least a 1080p display, Android Auto, WiFi, bluetooth and supports 4 camera inputs for a 360 degree systems like in modern cars. So I started to look for DIY options including a Raspberry Pi but the best OpenAuto OS for this usecase shut down recently. And considering the image processing limitations of the Raspberry Pi, I started wondering if any of this would be possible on a Jetson Nano. With its Ai prowess, it can even have blind spot monitoring and other safety features as well. Thanks for making all these videos!
This is a very interesting thought. I believe it would be possible at least with blind spot monitoring, and a car would have no problem powering the nano as well. Very interesting idea, I do have a relatively decent project car in the garage that I could attempt some of these things with, I just don't know if anyone would be interested haha.
@@Bijanbowen I appreciate you even thinking about the implementation of this absurd idea lol. People don't even think about going this route, is perhaps the reason no one has had talked about it yet. But when they get to know they can make the fastest, most powerful and feature packed car stereo, it will blow up.
@@ashyouknow7420 It is an awesome idea. I have been toying with how to go about using this to put blind spot monitoring in my 1988 Saab ever since reading your comment haha
@@Bijanbowen I'm honoured lol. Waiting for your video then👀
Hi can you try to use usb-cam instead of RP-cam? I get errors... thx
Sorry to hear. I see the usb camera input functionality is still in beta, but maybe this can provide some assistance, good luck and hope it gets solved! github.com/hailo-ai/hailo-rpi5-examples/blob/main/doc/basic-pipelines.md#usb-camera-input
I can’t find the Jetson Orin Nano for sale anywhere! 😢
Back ordered from spark fun got it in about two weeks.
I went ahead and ordered one on backorder. No idea when it will ship, but I'm hoping it will be in by February.
Hopefully stock issues are resolved soon!
Arrow has almost 300 of them.
@@nombable Awesome -- thanks!
Hey, good video…. How does Frigate compare between the two units?
Thank you! To be honest, I have not used frigate so I can't say for sure. Looking at the github for it it seems rather powerful and something worth investigating.
Thank you footage video , can I retrain the model ?
Yes, though it would be an in-depth undertaking and would need more powerful hardware than either of these boards.
i love your vedios
Thanks very much!
which is better Nvidiai jetson nano orion or RPI 5 with AI extension , like for transformer based computer vision , segementation , classification of realtime vision , and which camera module is suitable for it
Jetson will be "better", pi is easier to use but not as powerful. For the Pi you can use any of the rapsberry pi cameras and for the jetson it becomes more complicated and you pick depending on what your use case works with USB or CSI camera.
Is it possible to train new objects with either of these?
It would not be a good use of time as they would be quite limited in their ability to train if at all.
Thank for being interested in turning the Jetson into a gaming computer I was wondering if you could do it I would be interested in purchasing it if was possible
Thanks! That is just a gaming comp style case, unfortunately getting the jetson to actually game seems to be a rather difficult task that may not be worth it in the end.
Does the Jetson have a cooling system or any other add ons to it
It has a fan that seems to do a good job.l
Thanks, i just wanted to know what this thing was capable of
No problem!
what kinda camera is on the tripod thanks
It is the raspberry pi HQ camera with a telephoto lens.
Very cool example. I like Vit implementation but the VLM can do better right? Like describing the image. Im thinking to extend Frigate functionality or even write frigate inspired pipeline detection but do some improvement after CNN detection and pass to VLM. With Jetson nano, i think this is possible. Example use case as security camera , VLM can describe parcel delivery guy etc. if the small VLM listed on jetson website can return json , then can do many things with this. 100 local
Thanks very much! Yes, VLM can absolutely do more in terms of descriptive power and a VLM with a json function call to perform an action if X is detected would be very very cool and replace a lot of stupid subscription service home security cameras, even if it was just to send a text notification of movement or something like that.
With that big camera i found that i had to take the ring closest To the pcb off and then i was able to focus.
Thanks for this, I will try this.
Been trying to find the Jetson for that price
Check seeed studio as they are backordered but it is listed at msrp and not some rip off price.
How do you program these to send a message or email with a screen shot if it detects something
It is possible, honestly chatgpt/gemini/claude would be able to walk you through this
I have a homestead and thinking about setting up a Pi 5 with a camera running AI with Python. Then ask the AI if it sees an image of my dog chasing the chickens to then play audio noise of me saying NO or STOP!
This is a perfect use case for that device.
I can’t find any jetson nanos anymore
Hopefully they're back in stock soon!