Building a Camera That Can See Wifi | Part 3 SUCCESS!
HTML-код
- Опубликовано: 8 сен 2024
- What would it be like if our eyes could see more colors? We may never be able to answer that, but we can build devices that can do the work for us. Over the last three weeks we've been building a DIY radio telescope tuned to 2.44 Ghz which is right in the middle of the wifi band. With it we hoped to take a picture of a building and see all the wifi routers inside it. This is the third part of the series and I'm pleased to announce the project was a complete success! If you haven't seen the first and second videos, I'd highly recommend watching those first.
Part 1: • DIY Radio Telescope Ve...
Part 2: • The Angriest Radio Tel...
This week we mapped the inside of the hackerspace (Foulab) I've been working in and then stepped up to taking an image of the whole building. Not only were we able to see our router, we could see at least half a dozen others, and even one across a courtyard at least 100m away.
The images were very large since they were full 360 degree panoramas, so too get a better view of them head over to my facebook page at the following link to see them at full size!
Image 1 Inside the building: bit.ly/2LXmA92
Image 2 Full building: bit.ly/2JZMvfI
Want to try and build one of these yourself? Here's a link to all the code needed to run a robot like this and turn all the captured data into an image: github.com/FOU...
__________________________________________________________________
A special HUGE thank you to Foulab and especially Paul for making this project possible.
Foulab: foulab.org
__________________________________________________________________
Some more resources
Power meter GNU radio: www.tablix.org...
DC offset fix (Didn't actually end up using this, but it's still useful): www.rtl-sdr.co...
GNUradio: gnuradio.org
____________________________________________________________________
Support the show and future projects:
Patreon: / thethoughtemporium
Nebula: go.nebula.tv/t...
Ko-Fi: ko-fi.com/thou...
Become a member: / @thethoughtemporium
Store: thethoughtempo...
______________________________________________________
Our Social Media Pages:
Tiktok: / thethoughtemporium
Instagram: / thethoughtemporium
Facebook: / thethoughtemporium
Twitter: / emporiumthought
Website: thethoughtempor...
_____________________________________________________
Cuestion. Something like this could be used to find the cellphone signals of people buried in colapsed buildings???
Really cool idea sir!
Actually... probably yes. It would just have to scan a lot faster and scan all 10 bands of 4g lte (200 mhz, 400 mhz, 800 mhz. etc.)
@@LightCodeGaming We use to use a setup like RFID for finding radio transmitters, an antenna and receiver is a tuned circuit with a standing wave ratio, it will echo back some power when you apply a signal to them.
@@LightCodeGaming They can be found in some prisons and such, these devices are prone to false positives though, and it is a PITA to keep them up to date with the latest tech.
One major foil to these types of systems is that high freq singles respond poorly to things such as concrete and metal, metals can both reflect and absorb at that freq range and concrete tends to absorb as well, so it would be a very limited use device for such a thing.
Personally if I had to invest the coin for something for use in that, I would lean to an active RF Detector, yes I know there are edge cases where the phone may be powered off but still functional, but it is far more practical to detect a transmitting device then to hope for a lucky echo, that an a safety concern is that anytime you have a building collapse there is going to be many unknowns and it is plausible that an RF transmitter can cause other metals to spark, all it takes is a bit of coiled conductor with some oxidation that just happens to be making partial contact with a grounding surface.
what about buried in snow?
Im pretty sure this guy can see into the future at this point
The bread bois
didn’t expect to see bread boys here, but what a nice surprise
I'm surprised to see you guys here
Is this Dad or son talking?
Dad?
Mind blown.
I was actually having a conversation with my 11 year old son this morning, we were talking about the electromagnetic spectrum, during our conversation we both wondered what the world would look like if we could see the entire spectrum - as opposed to the limited visible portion that we currently see.
I get home and this video pops up in my recommended videos. Amazing 😂
yea its nearly impossible to even try to imagine what the world would look like if you could perceive the entire spectrum but definitely fun to think about, and this experiment was awesome
@@SpydersByte
100% agree.
@@vzgsxr If you saw in the Electromagnetic spectrum only, all you would see is likely a black and white gradient between high and low signal sources, because the electromagnetic spectrum does not reflect or interact with solid matter like the visible spectrum does. You would not be able to see anything physical, and would be blind to most perils around you. It's likely life could have developed with a way to detect this signal, but unlikely, because it would be relatively useless as a way of navigating the world around you... unless you wanted to find Wifi routers?
@@psilobomthe visible spectrum is the electromagnetic spectrum....you have no idea what you're talking about
It would be a confusing mess. There would be random lights coming through your walls from radio towers and cell towers and WiFi routers. There would be random lights coming from the sky in the form of xrays and gamma rays also. It would just be confusing. Most buildings and people would be transparent and nearly invisible.
A few months ago I was taking a physics class and the professor was talking about “what if we could see wi-fi waves” and that amazed me, it made me wonder how it would be like. To my surprise, I found this video and wow... thank you guys for making a “what if” a possibility!
what-if.xkcd.com/
Your phone is spying on you and sharing its info with your YT recommendation algorithm
@@jamsterical8467 my phone was in my backpack sealed :v but thanks for the warning
@@anthonyrivera8530 It can still hear you
@@anthonyrivera8530 ignore the consiricy theoryists just a coincidence nothing more it would take so much processing power to proccess all of that informaiton theres a reason why alex etc only are active when you say a code word
This is the coolest thing I have ever seen on RUclips. The Stackoverflow moment is too real.
wait for the karens
@@bighatman3572 you talkin bout kens and karens thinking wifi = cancer?
"For those of you interested in trying this yourself-"
Me: *cries in potato*
I’m here for you. 🥔
We understand it !
@@patricknelson very interesting
xDayWolf *cries in morse code*
lmao
Amazing. I am electrician and am i curently working on an internet project installing access points in one company. Yesterday i was thinking how can we build a cam to see signal and making things easier to install. This video was in my recommended today! I knew google were spies but i didnt knew they are reading mind also. This new black magic marketing is freaking me out :)
Google Analytics will smugly take credit for that ... lol
probably :D
LOL
Your phone is listening to you. So if you talked about this idea you had with someone, that's where they got it from.
@Dread Man The take whatever other unrelated things you're searching/browsing, and then make predictions of them. Kinda like that one girl who was pregnant, but didn't realize it, and then Target sent her ads for pregnant woman because their algorithm detected that she was pregnant since she had been buying scent-less lotion.
The code in line 68: print("Shit on fire, yo!"); LOL
We were on the verge of greatness, we were this close
How to tell when someone is a newbie programmer 😄
wat min ? i was listening the video xd
Guanete a bit after 1:43
@ 1:47 😂😂👍
I work for an ISP, we have an HFC network, which is hybrid fiber/coax.
You could get rid of a LOT of noise just by buying the (very cheap) tools to put a proper connector on and passing it all through a filter. That and using some nice TFC quadshield coax. It'll clean up the process so much.
All of that is really quite cheap.
NASA: Stay right where you are
Naysa is space
NAZA: Stay right where you are
Nah we already have waaay better radio telescopes
@@sirsanti8408 That's due to the fact that the dish that radio telescopes use are HUMONGOUS and can concentrate all the of it on one focal point. They managed to do one with a TINY receiver and coding. If not NASA, a private company will definitely want to hire these guys if they could.
shit's comming
If it wasn't already awesome enough that you completed such an intense project, you also gave us the links to do this project ourselves??
You guy are...AMAZING!
Is part of the big agenda
I'ts not already easy, because the signal bounces in many directions, and lost streng across the rocks and iron of the structures.
if you say so lol
+unixmonk Just Admit It, Quit Whining Like A Baby, You Couldn't Do What The Thought Emporium Had Done. X'D
Regarding the multithreading problem:
The python interpreter has a global interpreter lock.
So, while you’re having multiple threads, only one is calculating at any given point in time
@@danukerudesu8919 Why is the data set so large? Don't you just get one value in that position?
And why not just start building the image as the data comes in?
@@dmhzmxn It's likely that each point they scanned from took a bunch of data over a short period of time and they averaged that time out later so they could more consistent data.
damn VPN
@@Tony_Goat I wonder why they're not averaging the data as it comes in... That'd save some serious processing time.
@@Loebane It's usually good practice to separate the two steps. That allows them to write and debug them separately, so that if something goes wrong with the image construction half way through the scan, they don't waste all that scanning.
6:20 - The moment where you see what an actual developer does like 95% of his time while coding.
I swear that website is godsend
Yesss lol
At first I was like "how do you manage to fail at reading every line of a file"
Then that search gave me a hint at what the problem was (assuming it isn't B-roll) and yeah that's an understandable mistake I'd have made.
Stack Overflow is an actual blessing.
@@shaneclark8903 stack overflow is nothing without fellow programmers. The programmer community is the real blessing
i dont know how you guys have time ot do all of these.
Thanks to patreon, this is my job. Every bit of support helps me make more videos.
He doesn't have to work, he is a filthy commie :p... Anyways great video series. Keep it up!
Cody Slab hahaha
teiser real fact. Nerds are making the shit dope. Then shit gets wild and we buy it. How does things works? Fuck do I know but it's codes n shit, digits 0101. Stuff like that. These motherfuckers then get filthy rich because they smart and know weird shit and they say fuck the cyber 12. Gang shit.
A scientist will always find time to do something interesting
Awesome work guys love it. You should be be working on the Square Kilometre Array in Australia!
Hey it's Turnah!
Haloo
Sweet project dude! Just wondering could you have used the same satellite finder meter you used for your previous video to receive signals from this antenna? I know this is for 2.4 GHz and the satellite meters max out at about 2.1GHz, but if you had a 2GHz antenna on here, you think that would work?
Keep it up!!
Maybe but I doubt it. If it's out of range the effectiveness drops precipitously. Doesn't really matter about the antenna. More gain won't change the circuits sensitivity much.
Glad you liked the video :)
jlaservideo, suprised to see you here
JLaser be like wifi pew pew
meow banana
@@thethoughtemporium hi
Alright, ma'am so where is your router?
Old woman: I don't know what a router is?!
No problem! *takes out cogsworth*
1 hour later...points at device sitting on the coffee table.
_ "I didn't know that was a Router, I use it to heat up me drinks" _
4:10 can imagine being that neighbor.. "yeah so we invented something that can look through your walls and want to know if you could tell us how well it is working" 😅😅😅
*Next project:* using WiFi to see people through walls
Gonna need a bit faster robot :P
Already done, see
"Through-Wall Human Pose Estimation Using a Radio Signals" Mingmin Zhao et al.
Saleem Says doesn't work with wifi waves, but with radio waves. This invention already exists.
If people have their phones on them and are sitting still long enough, this contraption can see their phones.
It's easy turn FLEX CAPACITOR on by 234 to 0 257 on matrix
0:26 That's *geostationary.*
Geosynchronous are inclined to the equator, and do track over the Earth's surface.
The synchronous bit is they pass over the same spot at the same time each day.
Geosynchronous can also include elliptical orbits with a period of exactly 1 day also which results in a position oscillation in the East-West rather than North-South direction of an inclined circular orbit. Of course there are synchronous orbits that are both inclined and elliptical also which cause the satellite to move in an Analemma in the sky. Of course there are plenty of other orbits that are at least semi-synchronous and repeat at intervals too commonly used for activities like Earth observation when higher resolution is required.
You guys are both right, but you're correcting who called a square a rectangle: there's no point. Geostationary orbit is geosynchronous. Only nerds like us would ever care that it is specifically geostationary and, well, we already know that. ; )
Cole Smith screw that, I learned something! I've always only heard it referred to as geosynchronous, and now I know there's more to it than that. That's awesome! Keep correcting people so the few who will bother can continue to improve!
Beware ninjabongtoker1 is a Flat Earth moron.
This is quite fascinating...
I'm blocking you
Quite
Bro stop following me
He's trying to become Justin Y
This man is everywhere 🧐
That labour wave background 😍😍
ツ
So you’re saying
If we shoot this in space we can see alien WiFi, that’s dope .
Yus
Unless of course they use a different frequency for their wi-fi if they even have wi-fi
Isaac Wildflower I heard of a theory that certain frequencies may be an avenue of communication. Kind of makes sense in my small brain
We have actually been trying to do that for decades(not with a wifi detector tho).
Max gamer 20 maybe aliens have really strong WiFi, who knows
I've been googling this and looking for apps that do this forever,well at least for years. I'm so glad someone else thought of this and had the means and knowledge to take it from idea to function because I couldn't do it myself respect, ps I have many ideas
Man I wish I was this smart.
Smart people are sad. Dumb people are happy. Do you really want to be sad?
@@RyanUptonInnovator I'm dumb and I'm sad all the time :'(
alexander deluna you are lol
@@RyanUptonInnovator smart people are happy as hell.
@@RyanUptonInnovator
Low serotonin low testosterone beta males are sad. It has nothing to do with intelligence.
"The only true hacker space in Montreal" ... Dammnn ... Dude's cold AF, just shittin on fools.
How so? Is there some other hackerspace in Montreal that he's trying to imply isn't a true hackerspace?
@@rokstr222ify Montreal has a huge tech scene, from games to AI, so there are definitely others around the city.
OMG Montreal!!!! It's my dream to go there!!
@@Nesggy don't
@@user-yk5in5mz1d what?
8:40 I think you got a visit by the government after this idea 😄😄😄
that's intended.
He was never heard of again.
He works at Area 51 now
That’s more of a US government thing, not a Canadian government thing
Well actually he supposed to dissappear mysteriously...
Uhm..
There is a good chance that that second patch is your neighbor's router. However, keep in mind that directional antennas do have some gain behind them, so you may be seeing your own router again off the back of the antenna.
could you add a metal plate/grill/microwave door behind to stop that?
Man imagine what this could do with VR and realtime imaging
paintboy360 I was thinking the same thing. Imagine walking around your city with this.
That will be in future as augmented reality glasses you may see all waves around you ;) cant wait!
@@TeternalGIone we won't see kid trying to walk one square at the time but dance beteween invisible waves
Isn’t this similar to what Batman did to find the Joker in The Dark Knight
@@black_swanN true dat
Having been interested in communication's from an early age, I came up with the thought that if someone developed a machine capable of actually seeing R.F., that we very likely wouldn't be able to see our hand in front of our faces, due to how many signal's there are coming from Radio, Television, Cell Tower's etc. !
I ventured too that if nothing else, someone using such a device, would perhaps assign a color for the various radio frequencies !
Looks like these guys are getting closer to that end !
It would be very exciting to be able to visualize a radio signal, as it leaves an antenna, and instead of relying on metered measurements to try and understand the antenna pattern and gain direction, we at last could see them !
Perhaps some day ?
🤔.....📡 ???
This is so inspiring to watch. Just two dudes exploring the possibilites and having a fun time and in the meantime contributing to something meaningful! Keep it up guys
You did it... Results are outstanding....
I did some research 20years ago with something called celldar, it was a system to generate a realtime map of areas using reflected RF from phone towers in the UK, it could also act as a passive radar system. It never got beyond the stage of "shapes" moving as computers were not that fast at the time. You could monitor an area and see what was moving such as people and vehicles (or shapes of them).
It was no problem with monitoring aircraft overhead, but that was limited by downward tilt in antenna radiation patterns and lack of reflection of 2G and later 3G RF.
you could run a fiber optic cable up the center to a lens at the tip and capture the visible spectrum to overlay it.
There are already cheap devices that do exactly this. They are called cameras.
Lmao
@@damianreaves7502 no imagination.
No imagination because ive already reverse engineered the human brain and discovered the full secret behind how you third dimensional beings operate on planet earth.
@@damianreaves7502 well you are either telling the truth or you are just a lame Troll on the internet with nothing better to do than object to every interesting idea that you are envious of because you don't have any thing to contribute.
Now this is a great project with a practical application. These ones are gold.
I just have one thing to say...
Binary!
As in, instead of storing the values returned from the scan in ascii, you store them as binary.
Depending on the maximum resolution of the data sent back as floating point values from the scanner, you could get away with sending them as single byte (0-255 integer), half-precision float of two bytes (float magic) or single precision float of four bytes (same float magic but better)
Either way, every line in that data file you are saving...
xx.zzzzzzzzzz = A total of 13 bytes
instead of
WXYZ = A total of 4 bytes.
And looking at the few examples you have visible in the video, most of the values seem to range from 30.X to 50.X so you probably could get away with simply converting them to half-precision or even single byte readouts at the source.
You should have ran a test to see just how many different values you get in the readouts and saved you a lot of dataset storage size as well as post-processing time. Perhaps even getting rid of post processing completely.
FBI: Cogsworth is coming with me. It 's a matter of national security.
They cant actually do that
No, they can't. Montreal is out of their jurisdiction.
you don't think they have pull?
In order to apprehend people abroad, the FBI requires consent from the host country, and congress has to grant them extraterritorial jurisdiction for said mission. The threat would have to be very credible, and very serious, in order for both of these parties to cooperate.
Bottom line: no, they don't have "pull".
And we're running with it. :)
I suggest to rewrite the program - run the averaging during reposition - you will save the postprocessing time.
My idea exactly! (I scanned the comments first to see if it wasn't someone else calling it before.) In 250ms I believe you can do more than averaging it. You can also compute the final image pixel by pixel. Also, why is so much data? How was it stored? As text? (also @The Thought Emporium)
I understand that for the prototyping part data is important. Especially when you are not sure what to expect. I was implying that once you know that averaging it is the way to go, you might do it on the fly. Very cool project btw!
Gigabytes of data is just insane. Back of the envolpe calculation:
At 3:03 you can see that the image resolution is 80 by 270, thats 21600 samples
each sample is a float, thats 4 bytes, or 8 bytes if you are using double precision (which I think python does). That means the robot should sample about ~168 KB of data. HOW DID YOU INCREASE THIS BY A FACTOR OF ABOUT A TEN THOUSAND??????????
The second time, it took 4 hours, with 4 measurements per second (at best) that gives 57600 samples. Did you include a Harry Potter book in each of your pixels? 44 gigabytes is ridiculous. (Edit: with each pixel 4 MB and 44GB of data, one would actually get around 11000 samples)
What is in those files that they take up 4 MB of space?
Nice project though.
During development it's more important to collect as much raw data as you can get, disk space usage and processing time for analyzing the data isn't much of an issue: If you have bugs in your processing code (like the wrong averaging mentioned in the vid) you could just rerun the analysis. If you throw away raw data and just keep the (possibly flawed) results, you need to redo the whole scanning....
I disagree about collecting as much raw data as you can get during dev. You need to choose sample rates wisely, because rather then helping development, if you have so much data that processing takes minutes or hours per pass, you could be wasting time. If it takes you a couple of passes to work out a bug or find a way to make better sense of data, having the right sample rate could make all the difference in a days work.
Creativity , innovation and dedication , intelligence are on next level 🥺🥺🥺🥺
...only intelligence agencies...
What an inspiration to all
This is truly what makes expanding your mind can do and create
It’s not only what you see but what you cannot see that makes life so worth living Great work bro👍🏻
It's 3 am, New Year is today. AYY YO What about that robot that can see wifi?
It's not new technology
Huge props to Paul, this project sure wouldn’t be alive without him!
At the end of the code it said “Shit on fire yo!” As an exception function string XD
That's dope that you two were willing to share your work and help others who would be interested in trying this experiment!
Good lookin
6:26 nothing defeats the good ol' stackoverflow
2:47
! L A B O R W A V E !
based af
This was an awesome project to watch you complete! Great job guys!
I think you will get far better resolution, using a very low aperture, high front to back ratio antenna, instead of a high gain and lower FtoB ratio antenna like the helical you are using in this.
A waveguide feedhorn is what is usually used for this sort of thing, or at a pinch a Pringles can.
You lose gain, but you gain in resolution and directionality.
put the antennae in a Pringles tube (tinfoil covered tube ) the longer the better , . It will reduce off center signals
We wanted to but for this it would've thrown the weight off and I didn't want to mess with it once it was working. We'll use that for the next version. Thanks
there are some antenna shapes out there you could use that would solve the hole counter balance problem " patch and Biquad " biquad is probably the simplest to make .if you change to 5.8ghz the Ant's are half the size for just as much gain fyi . still need to be shrouded ......ill stop, there is a shit ton to think about with RF and waves in general
The Thought Emporium if you get serious about trying to narrow your beamwidth you should really look into the monopulse technique. Check Wikipedia. Side note I would love to see you add a transmitter and take radar pictures of things.
The Thought Emporium . Pringles cantennas are not the right diameter for wifi. There's online calculators for diameter and cutoff frequencies and wave guide can feeds for dishes are fairly common. Cool project!
i need to read up on that more. Sounds very interesting,
Why wasn't this recommended to me sooner?!?!
The things a couple of genius brains come up with. It's so awesome. I'm in awe. ❤
"Laborwave" I love it
Bravo, Gentlemen, bravo! You, Sirs, are badasses! Keep up the fantastic work.
A hand-held wi-fi mapper would be quite a commercial success in music shops. Thousands of musicians use wireless links, mostly at 2.4 g.
If anyone finds this interesting you can get started with some RTL-SDR’s for $10 and view the a 25 mhz to 1.8 ghz spectrum using SDR sharp.
Purchase Protection Insurance applies to covered items for ninety (90) days
(other than in the circumstances listed below) after the date that the first
part of the full purchase price (or the full purchase price) of the covered item
is charged to your card account.
Cover will end at the time any of the following occurs during the ninety (90)
days after the first part of the full purchase price (or the full purchase price)
of your covered item was charged to your card account:
• your card account terminates; or
• you cease to be a resident of Australia; or
• you become bankrupt or enter into an arrangement with your
creditors under the provisions of the Bankruptcy Act 1966 (Cth).
I wish I knew what any of that meant. Super cool though
8:26 if you pause the video there and wiggle your phone around (if you're watching it on there) the plane wiggles :)
How to check if someone's broke by judging their wifi strength
I think to do that you can-
Step 1. First check the strength of the signal when your average number of devices are on.
Step 2. Add one more device and now check the strength of the signal. Subtract the value from the average load.
Step 3. Now if you scan a drop in signal strength by the value you noted in step 2 or more than that, and the drop is originating from an unknown source then you can determine there is a break-in and from the magnitude of drop you can calculate the average number of unwanted device being used.
what if they are just stealing your wifi?
Smuggerino they should call this a “drip checker”
....or actually use it for something useful and save the lives of all the people hopelessly trapped under marinara sauce. Millions every year die in the sauce because we simply cant locate them in the hardy sweet and tangy sauce.
@@the_original_Bilb_Ono what
"I'm a turnip" is the absolute least toxic way to have a self-depricating sense of humor that I've ever heard!
How are YOU not rich or famous?? This is the stuff our world should be on about instead of processed foods science and genetically modified diabetes. Congrats on accomplishing your project. I find this awesome and a your stuff is so interesting.
I remember reading about a guy who was super into body modifications. He got some sort of implant on/in his ears (i think) that allowed him to hear wifi and stuff like that
Laborwave on thonkpad. Nice
More like THICCPAD (X220 ftw)
Watching on a T410
Watching on an X201
I need this so bad! Then I can finally prove to my parents that the wi fi signal is weakest on my bed
amazing work, well done - so impressive and you're so generous with your learning.
You know what? Rotate/sweep a reflector in front of the antenna instead of the whole antenna.
This could get the scan faster and snag wires no more.
I don't know if you're familiar with the early versions of ultrasound (echograph) probes that swept a reflector in front of a single transducer. It would limit the field of view at around 45 degrees but heck, 360 cameras are recent tech too :D
Also the focusing could be improved like this by shielding the antenna better. The helix has ugly side-lobes and it will give plenty false readings. A very long can antenna (2...3...4 lambda wave guide) has a narrower beam pattern and no side-lobes.
Isn't .25s a bit long?
The "how do I change the number of open files in Linux" is that part that got me.. knowing Linux is always a work in progresss and if you're worried about open files then you've got a lot open.. showing the work you put into this is awe inspiring, not to mention the results are amazing.
My type of guy is one who aspires to build machines to see cosmic hydrogen and pulsars. How can that not get your engines revved up? ;)
Liked the video for the name cogsworth
This channel should be the single biggest channel on RUclips
Really cool idea. I cant help but think it would actually be better in black and white though as that way wifi signals will be bright spots with lower signal areas being darker. Colour makes it harder to interpet?
also the colors fluctuate a lot into completely different colors; is that a bug in color coding?
Guys! That’s amazing! Thanks for sharing all your work!
I'll be totally honest. Paul looks like a guy who helps run a hackerspace in Montreal. Couldn't have casted it better.
Yay, now I can see if my router is screwing up again before my computer loses the Internet.
I call it Starbucks customer.
But you see very little
I always wonder how people like these guys make a living doing cool projects like this?
K, now miniturize the scanner and deploy a swarm of drones to automate building scanning.
'miniturize'
At around 2:40 mark you talk about scaling the values (floating point numbers between 24(Minimum) to 45(Maximum)). I think you should revisit the formula. The scaling formula usually reads like (Value - minimum)/(Maximum-Minimum) * 255. This will give you better separation between values. Or am I missing something here? Awesome work though! You should also explore using GPUs to cut down on the processing time.
Remember that the reading is of Decibels, which are not linear in progression. Decibels are logarithmic. Matching the test value to what can be expected to be output and measured is a crucial factor though. I'd agree with you on sentiment if not the exact method you propose.
I agree with you. He's normalizing the value using 100 as a max. Which only works because you're not going to get a value over 100 dB, but you're not getting the best separation between the value. Also, for the image to actually be colorful, he's not actually using the same value for each of the RGB channels, or you'd get grayscale values (which the resulting image clearly is not)
I can't really understand the code that generates the image to know how the normalized values are being converted to RGB though.
Edit: Under more investigation, it seems that the code's output is indeed just a 270x80 grayscale RGB image which they're later converting (and upscaling it, which causes artifacts visible in the final image) to HSV, just so it's colorful. Personally I'd map the intesity values to a gradient of colors from blue to red, creating a more normal heatmap, rather than the trippy colors you get from mapping RGB directly to HSV.
@@esuelle well, I guess maybe if you can just work out the central area of each of the cluster groups, and paint out the highest decibel level areas, you'd get something meaningful from the data. A lot of those are going to be reflections from the WiFi hotpots but they should centre roughly over the position of the routers sending the data. The tricky part is a computer working out how many clusters, how many routers working within the scan area.
@@esuelle Thank you, that is exactly what I was curious about. Same number for each RGB == grayscale; HSV makes sense. Asking the real questions.
"Labor wave" always nice when your favourite youtuber is also a commrade
Saving to different files have lots of cons. You have to write and read it again in 2 phases. That's already the double of io operations. I would rather do the stuff in other way.
1st idea: create a fs partition that allows huge files. And save only one file and seek data using different pointers. You only need different file descriptors set as readonly to get this working and little work to not read what is not written yet
(((1.5st idea: you could use tmpfs a lots of ram too)))
2nd idea (which i think is better) is to pipeline everything not saving the file but creating buffers on ram and processing into image.
This could be very useful to plug several functionalities together. With python you could any functional-like approach or punch lots of objects with same interface.
Anyway, this project is so could that i would like to live somewhere near just to see it personally
Just tobe more specific io for hdd just stacks lots of syscalls from both linux vfs and hdd driver. Also depending on your fs settings it also pushes lots of meta data around. These stuff are not bad for general purpose because it gives lots of control and flexibility. But I think you dont need it. So avoiding the overhead and pipelining the process would make the project lighter im terms of requirements and easier.
But I dont have any cool projects like these XD getting it to work is totally cool!
Danukeru DESU :D cool a reply :D @me in rd stuff too :D s you could have both. Streamed to send data and stored. Or streamed reduced and processed than stored. Anyway, while i was watching i saw you guys looking foe how to increase the number of files. So maybe protobuf oe anything could helped too. Anyway, im not anyone important :D i mean, YT comments who cares :D great work anyway!
I wonder if one could do this with a usb wifi stick with a detachable antenna. Run a scan, move the robot, run a scan, move...
Then you could color in different wifi routers. Or colorise by channel.
That was similair to the original plan actually. We just couldn't figure out how to implement it in GNUradio since we're new to it.
You would have custom code for that, no GNU radio required, which wouldn‘t be so hard, I‘ve done it few years ago for just 1 axis of rotation and the produced some nice polar plots. I can‘t quite recall how it was called, but there is a really nice library for python, something like „scapy“ or similar, I am sure you would find it. Of course if you‘d run it off of a raspberry pi or similar you could do the turning, data collection and processing all from within the same device and even code :)
a Wifi-Stick like the TL-WN722N (with Atheros chipset), set the thing into "monitor" mode, and then use something like pcap (or it's python equivalent) to capture all the packets you can receive.
Filter out every SSID announcement packet and plot their RSSI onto an image. That should look really cool and because you have the mac address of each router, you could even give each router on the imageplot it's distinct color...
Yes it is in fact doable. Just with the strength of the signal. I did a similar experiment when i were in Montréal. At the time i had my Wifi attacked all the time. I got fed up and i made an honeypot, when the guy with the same MAC connected on it, i went with my laptop and a cheap wifi on promiscuous mode, gathering the data while walking, to then pin point it in an apartment nearby using triangulation.
MalakiLab Then what?!? How does this awesome story end?!?
I may have commented on this before. And I know its old. But this is just like the scene in Continuum where she uses the distortion of radio waves to see what happened at a crime scene. This was in S1E7 of the show at around the 27/28min mark. This is pretty much the same thing. This really is fascinating tech.
*(hackerman intensifies)*
All my wifi is 5ghz tho, haha can't take a picture of me.
Love your stuff mate.
a small amount of tuning would allow him to do so
Like whole new receiver :D
He can tune it to 5Ghz if he wants.
A few months ago, my friend sent me a LABORWAVE desktop wallpaper and now I see the exact same one in a random video on RUclips.
When a video cuts out and you try to find the spot where the best connection is
Try using one of those old radio tv antenna dishes, for your next antenna
Used to have one of these tuned to 2.4Ghz, and I stupidly got rid of it.
he did in the previous version
tesseract, that was a satellite dish, not a radio dish
true sorry, that would be a huge feat though
tesseract, it would be a huge feat, but if he does it and I works....
Why do I think it would be so cool if they could do a phased array? 8:45 Oops, posted too soon. Yeah, definitely want to see that.
Ahh good ol stack overflow
Hey there! That's a truly amazing project. I have a suggestion: what if the robot head/antenna could spin continuously in both axis, without needing to reverse the rotation? That would enable 360-degree mapping of the surroundings. I hope that makes sense.
P.S. Also, do you guys know if a simple adjustable class-c rf amplifier would do the job?
Make a mechanism similar like LIDAR. The antenna may be coupled by rotational trafo or capacitive to need no sliding contact.
A fellow in Oregon was working on a similar system to overlay on a Video Camera back in about 1996, saw some samples of his, but not sure what happened to his project. Yours would be interesting if it could automatically do the same thing, as you were doing with photoshop, merge the still image with the emissions. Possibly generating that at video frame rate for video output, similar to Flir heat camera. Excellent work, thanks for sharing.
It could be worth looking at the numpy python module to see if it can do a faster job processing the data.
I think you could load to data directly using numpy.fromfile. I would test it out, but there is no sample data on the GitHub page :P
I also think that numpy can help to do a faster job. However, i'm prety sure that numpy.fromfile is not suitable for large amount of data. In this particular case, i would recommend using generators to read and process data without put them in RAM.
I don't know this functionality but it can be interesting. If it's possible (and it's not always possible), I think it's better to do all the processing on the fly though.
If you really want faster processing of huge amounts of data, may I recommend not using an interpreted language to do that processing?
You should try making it capable of higher resolution
Now I can prove to my friends that my WiFi isn’t bad
NASA be like “shut up and take my money”
that would be a first
My small brain can’t comprehend this but it’s really interesting 🤔 and cool
I have zero experience with all of this but it looks like so much fun!
1:38
Bruder muss los
Bruder muss groß💩
Bruder lass los
Awesome project! If the first image processing only used the first capture, do you really need 250ms per location? Seems possible to speed up the sweep and lower capture time to ease data transfer and post processing burden.
You would want to sample over a period of time at each position because 802.11 transmissions aren't constant over time. They start and stop, and there's often lots of dead time with no transmissions followed by periods with many transmissions. If you only sampled a position for a short amount of time, you might miss a transmission.
In one of your experiments try building a faraday cage around one of the rooms your camera sweeps over and see what it looks like.
Can you do the same scans for 5G to see howmuch more powerful it is?
5GHz WiFi has less range than 2.4GHz and is blocked more easily by walls.
@ same as wifi
@@infinitedeathloop5517 no...