Raspberry Pi Lidar Scanner
144 comments
·April 19, 2025godelski
Aurornis
Learning projects like this is about to get a lot less accessible due to the extreme tariffs and elimination of the de minimis exemption. Take that BOM and multiply it by 2X or 3X depending on the source and how many different shipments arrived.
I can’t tell you how depressing it is to go from having access to cheap learning materials for introducing kids (and adults) to electronics, and now it’s being taxed away in the name of improving the US competitiveness or something. Total footgun.
godelski
Doesn't change the fact that the advice is still beneficial. At worst you still have a good history of the effect of these tariffs.
I'd call the tariffs the second death of hardware though. The first was when we killed all the parts stores. That was a slower death, coupled with the loss of right to repair. But we've been making big strides in that domain, so I hope we can undo that death. If we can also undo the dumb tariffs too then ironically we might have a chance to bring back hardware which somewhat seems inline with what that party (claims to/pretends to) wants.
macintux
A new Radio Shack dealer just opened in a rural area, with an emphasis on radio (it's near an off-roading park), and your comment reminded me what terrible timing this is for them. Such a cruel twist.
robocat
> we killed all the parts stores
We ran out of people buying from parts stores - hobby electronics became less popular.
Eras of hobbies:
Mechanical
Chemical
Electronic
Computer hardware
Computer software
[What's now?]helpfulContrib
[dead]
firesteelrain
The good thing it’s on GitHub so you can submit a pull request for a BOM to help the person out.
a_subsystem
This is the most ungrateful comment I've read today, harping away about how 'it should have been done'.
Well you fucking do it then.
I know that my time is so short (because I have a family) that if I can even do a project then I'm almost certainly not going to document it because getting it done will be enough of a stretch for me, and if I need to come back and re-do it again, I am probably not going to even bother. Not all of us live in mom's basement and have the luxury of extra time.
uxcolumbo
It was not ungrateful.
It was a general suggestion for everyone doing hardware projects and OP did a lookup and provided the additional info / links, which sparked further discussions.
Chill.
taylorportman
He did 'do it', and saved us all the 10-15minutes it took.
juujian
Incredible that this is too expense for a company like Tesla.
nisa
What is the HN opinion on Tesla skipping lidar? Having spent some time with computer vision in university I think it's insane to skip it - sure stereo reconstruction is powerful but lighting conditions have such an impact on results that having some robust depth data feels like a no-brainer and skipping it feels like malignant neglect.
godelski
As someone who's done a lot of computer vision, it is insane to skip it. And it's sad because what everyone missed from that viral Mark Rober video [0] was not the Looney Toons wall hit but the fucking kid in the smoke. Add all the cameras and AI you want, you ain't changing the laws of physics: visible light doesn't penetrate smoke. But radar does. Every (traditional) engineer knows that safe systems have redundancy. That safe systems have redundancy through differing modalities. Use cameras, but also use radar, lidar, and even millimeter wave. Using just cameras isn't just tying one hand behind your back, it's shooting yourself in the kneecap afterwards
nfriedly
> What is the HN opinion on Tesla skipping lidar?
Short-sighted and egotistical.
There likely have been deaths and injuries that would have been prevented by lidar, and there will likely be more in the future.
ruined
interesting claim i read in another thread a couple weeks ago:
>Tesla Vision is, currently, legally below minimum human vision requirements and has historically been sold despite being nearly legally blind.
zymhan
> the HN opinion
I'm not sure why you'd think HN has a monolithic opinion, this is a site with myriad different users.
teleforce
There will be time in the very near future (read five years time) people will not buy vehicle (car, bike, etc) without lidar as the price become insignificant as car reverse camera, and it become commonplace.
Personally now I'll not buy any vehicle without assisted camera parking and apparently many people will agree with this important feature including Marques Brownlee [1].
[1] Reviewing my First Car: Toyota Camry Hybrid! [video]:
xorcist
A similar situation to Jobs reciting research on how optimal the one button mouse is. A thought bubble.
Why, we will perhaps never know. But likely they were early and it was deemed too expensive back then, or didn't find a supplier they could work with. Now there's too much prestige in it and they can never back down which would be admitting to a mistake.
It would be one thing if it was a one time event but then they repeated that playbook with the lack of a rain sensor.
MOARDONGZPLZ
I can only speak for myself, but I work on this stuff in this industry: Tesla’s choice is asinine at this point. It’s one thing to claim cameras only and find that won’t work and pivot, but they are so dug in that they can’t admit they were wrong and won’t do so. So it’s asinine now.
colechristensen
I think it was a valid decision that turned out to be incorrect and is staying put as a result of stubbornness. People really like criticizing decisions in hindsight especially here where the armchair engineer with the benefit of hindsight is too common.
avsegal
This isn't too expensive for Tesla, it's just nowhere near the level needed for an AV. Automotive lidars are 10-20 scans/second, rated for dust/rain/etc, and need a range of at least 50 meters, but 100-200 is more ideal. Not a fan of Tesla's approach, but I wanted to clarify that it's not like they can just use a lidar like this and call it a day. The specs are completely different and that really drives up cost!
ndsipa_pomu
Maybe something like this: https://linuxgizmos.com/sony-introduces-as-dt1-described-as-...
> With a horizontal field of view of 30° or more and 576 ranging points (24 x 24), the sensor supports a frame rate of 30 fps, with a reduced 15 fps mode for maximum distance operation.
derac
The requirements for electronics in a car are pretty extreme (temp, durability), not that I disagree, but it's not apples to oranges.
lucb1e
> requirements for electronics in a car are pretty extreme
+ the salaries of everyone working on that stuff, not just assembly but also writing the code to support it
Not that I disagree, either: at the volumes that a modest car company puts out, I'd assume it's easily worth the, say, 3% cost premium on the car's total price to have something that can actually see things you don't see and thus makes a safer system. It might even reduce costs by having lower requirements for the vision hardware and software, but that's not something I can know. There's a lot of unknowns here that I think mean we can't really do a good comparison indeed
cosmie
That hasn’t stopped Tesla before. They have a track record of treating automotive-grade quality standards as optional when doing electronics sourcing[1].
As the article notes, Tesla conveniently “fixed” their thermals and durability issue that caused by inventing a feature called cabin overheat protection and marketing it as for people/animals overheating and not for the non-automotive-spec electronics in the cabin.
If you can’t bring auto quality electronics to the car, just change the car so it avoids standard auto thermal conditions ¯\_(ツ)_/¯
https://www.thedrive.com/tech/27989/teslas-screen-saga-shows...
qgin
They wanted to sell “self driving ready” packages 10 years ago, when LiDAR actually was expensive. So at the time, they had to make big deal about LiDAR being unnecessary.
mcny
But now it has come down in price reportedly by more than a factor of ten so at some point a logical person would revisit that decision
theoryofx
The lidars used on self-driving vehicles are far more capable and far more expensive.
xbmcuser
Not by that much current generation hardware for cars is $500-700. And some of the oem expect to bring it price down below $200 with the next generation equipment. Now that BYD put self driving in almost every car it will supercharge adoption and lidar prices might drop even a lot faster with economies of scale.
jsiepkes
At this point having "something" would probably even beat having nothing.
I guess it's simply a big numbers thing. If you sell lots of cars, shaving a couple of hundred dollars of each car adds up.
dreamcompiler
Karpathy addressed this question at the time:
https://news.ycombinator.com/item?id=33397093
Of course he was working for Tesla back then. His opinions might be different today given that Elon is no longer signing his paycheck.
godelski
What a weird argument by Karpathy. He has a degree in physics. How does this dude not know that radar can see things not possible through camera vision. That argument there doesn't make any sense. That there's supply chains and things break and this makes it unsafe? Well that's true for every single bolt, every nut, ever little thing. I understand a drive to simplicity but you can't just throw fancy words in there like entropy while ignoring the literal physics that says camera + radar is less entropy than camera no radar. There is literally more (unique!) information available to you!
jdminhbg
His opinions aren’t much different in interviews I’ve heard since, although of course that doesn’t mean he’s completely unbiased now.
firesteelrain
Do we really need LiDAR in a Tesla? I own a Chevy Trax and it has LKAS and ADAS. Not even using LiDAR just sensor fusion with camera and radar. It’s a cheap car too. It’s car assisted driving.
I have driven a Tesla once but not with the added feature.
Animats
The actual scanners: [1]
Max range 12 meters. That's when it seems to start to get expensive. The light source, filters, and sensors all have to get better.
Good enough for most small robots. Maybe good enough for the minor sensors on self-driving cars, the ones that cover the vehicle perimeter so kids and dogs are reliably sensed. The big long-range LIDAR up top is still hard.
gmueckl
I'd like to know where this price jump really comes from. Google doesn't help me. My first guess is that laser safety becomes an active control process at this point - laser scanner mirror needs to keep moving to not be able to deposit a damaging amount of energy onto a human retina. So you need a safety critical control system to constantly monitor mirror speed and and position and shut down the laser when it becomes too slow. How wrong am I?
Animats
More output power, larger optics, more sensitive detectors, more rejection of unwanted light, more pixels, larger rotating machinery, active stabilization... And the big units are low volume.
Here's a top of car LIDAR you can buy for about US$27,000.[1] 128 pixels high sensor, spinning. This is roughly comparable to Waymo's sensor.
amelius
Somewhat related. I'm looking for a cheap way to measure distances to approx 10 microns accuracy, over distances on the order of 300mm. Any ideas?
shellfishgene
Maybe check out this for ideas? https://youtu.be/qMYBwbTIL-0
bobmcnamara
Does the interval you're measuring move around much?
Can the measurement system touch or be affixed to it?
Sounds like a pair of nice calipers might work. So depending on your precision needs, you might get away with the same approach: sliding grid of capacitive cells that slide over the measurement cells. Microcontroller measures them as it slides through. Atan2() for the final result. The meter only part of this is called a DRO(Digital ReadOut)
aeonik
I have some design ideas for a diy system, how much money/time are you willing to spend for experimentation?
What counts as cheap to you?
I'm thinking about automating something a long these lines:
https://youtu.be/hnHjrz_inQU?si=dNzXVBVFsr7e8m_6
Off the shelf lasers and camera sensors can be hacked around with DIY for some pretty unexpected precision.
sazimi13
Thanks for sharing this video, I am also interested in this exact thing. However from my understanding with an approach like this you are limited by the size of the image sensor, meaning if my surface has a bump that is larger than the size of the image sensor it would not get measured. Any idea on how to make something like this work if the goal was to measure slightly larger topographical changes at a less granular resolution like in the 100mm range?
aeonik
Yea I have some ideas, but I haven't found an easy way to implement it yet.
The term of art that I'm exploring is called "Holographic Interferometry".
Sibling poster gave you a link to regular interferometry.
But basically, if you split the laser beam, one goes straight into the camera sensor, and the other off your object, you can do some pretty amazing things. Depending on a lot of little details (The devil is hiding here).
I found 3Blue1Brown's explanation to be the best, but less "ready to use".
Joel_Mckay
Making a fringe counter out of a Michelson interferometer is a classic project:
https://www.youtube.com/watch?v=j-u3IEgcTiQ
https://www.youtube.com/watch?v=ucuVsReDze0
https://en.wikipedia.org/wiki/Michelson_interferometer
Note those 532nm green DPSS lasers are repeatable within +-1nm across their normal operating temperature. Adding a 20nm wide OD6 narrow band-pass filter to a $5 5mW DPSS laser module is the cheapest precision money can buy these days.
Really depends what one can get away with in the mechanism being built. Note, many machines will fall under export restriction, and as a company people have to decide whether that encumbrance is worth the hassle.
Best of luck =3
djoldman
For what purpose?
RossBencina
Answering their question would be more helpful here, even if it doesn't solve their problem.
aeonik
Not OP but I'm in the same market, 3d printing and desktop CNC for me.
hamburglar
Assuming the XY problem based on nothing is pointless and counterproductive and only serves to make you feel smart.
Joel_Mckay
We had a similar issue at one point, and had to build something custom that cost way more than I'd like to admit. Thus, I would recommend just looking at DRO kits for CNC milling machines.
If your project is not budget constrained, than there are complete closed-loop stage solutions around:
Best of luck, and prepare yourself for sticker shock... lol =3
null
mannyv
You could probably harvest these from robot vacuums on ebay/goodwill.
mannyv
These = lidar sensors
politelemon
The sketchfab examples are fantastic, to be able to move around in a 3D space, like it's some kind of scifi simulation.
The mouse controls are confusing the heck out of me. It shows a 'grab' icon but nothing about it grabs as the movement direction is the opposite, feels completely unnatural.
heljara
I've been toying with photogrammetry a little bit lately, specifically for scanning indoor rooms and spaces. So far I'm finding metashape the most suitable for it, but some of the precision isn't great (but I'm still improving my technique). I mostly want to convert the interior of one real building into a digital model for preservation and analysis. I've briefly considered LIDAR, but put it in the too hard/expensive bucket. This project seems to challenge that assumption.
What does the software post-processing look like for this? Can I get a point cloud that I can then merge with other data (like DSLR photographs for texturing)?
I see in their second image[1] some of the wall is not scanned as it was blocked by a hanging lamp, and possibly the LIDAR could not see over the top of the couch either. Can I merge two (or more) point clouds to see around objects and corners? Will software be able to self-align common walls/points to identify its in the same physical room, or will that require some jiggery-pokery? Is there a LIDAR equivalent of coded targets or ARTags[0]? Would this scale to multiple rooms?
Is this even worth considering, or will it be more hassle than its worth compared to well-done photogrammetry?
(Apologies for the peak-of-mount-stupid questions, I don't know what I don't know)
0: https://en.wikipedia.org/wiki/ARTag 1: https://github.com/PiLiDAR/PiLiDAR/raw/main/images/interior....
rsp1984
Shameless plug, but if you own an iPhone pro or iPad Pro (which have Lidar integrated), you should give Dot3D a try. It does everything you describe and we made it very easy to use.
abraae
There's a lot of stuff that was better in the "good old days".
But to be alive when it's possible for gifted individuals to create technology like this is just incredible.
itissid
For home improvement projects, This could be quite useful for generating point cloud map of places hard to get to. Like I have drywall installations I would love to get behind and check how things look, this would be great for that.
chneu
This is a very legit and good idea. A simple stud-finder like tool to map out behind walls would be incredibly useful for folks who run cabling or whatnot.
Aspos
GY-521 in particular and MPU6050 in general make quite poor IMUs. Why do you use them? And what for in this particular case? What do they do in this set up?
HALtheWise
Do you have other sensors in the same price range that you'd recommend instead for most uses? How much accuracy improvement would you expect?
badmonster
Hi! Thanks for sharing this amazing work. I’m curious about the scalability and performance of PiLiDAR when deployed on large-scale outdoor datasets. Have you benchmarked it on datasets like SemanticKITTI or nuScenes? If so, could you share any insights on runtime, memory usage, and how well it generalizes beyond the indoor scenes used in your paper?
nisa
I think you (or me, please correct me if that's the case) misunderstood something here - this is a diy lidar scanner for data acquisition - these datasets are mostly created using rgba cameras and the point clouds are later created with some post processing step.
So it's not a model for processing data but rather a hardware hack for having a real lidar - as in real depth data.
You can throw anything you like on it.
donatj
Oh hey! This is exactly what I was looking for just a couple weeks ago! I've had parts to prototype something roughly equivalent to this sitting in my cart on Amazon for a couple weeks now, but I've been very uncertain on my choice of actual lidar scanner.
I'll have to look into this as a starting point I get back from Easter vacation
mannyv
How do you make it so your LIDAR doesn't interfere with someone else's LIDAR?
frainfreeze
Multi path errors are rare, and if it happens there are mitigation techniques.
This is really cool
One thing I'd suggest, for any hardware product, is that when doing your bill of materials to provide links and show estimated costs. Sure, these will change but having a rough idea of the costs is really helpful, especially when perusing on from things like HN. It can be a big difference for someone to decide if they want to try it on their own or not. It is the ballpark figures that matter, not the specifics.
You did all that research, write it down. If for no one but yourself! Providing links is highly helpful because names can be funky and helps people (including your future self) know if this is the same thing or not. It's always noisy, but these things reduce noise. Importantly, they take no time while you're doing the project (you literally bought the parts, so you have the link and the price). It saves yourself a lot of hassle, not just for others. Document because no one remembers anything after a few days or weeks. It takes 10 seconds to write it down and 30 minutes to do the thing all over again, so be lazy and document. I think this is one of the biggest lessons I learned when I started as an engineer. You save yourself so much time. You just got to fight that dumb part in your head that is trying to convince you that it doesn't save time. (Same with documenting code[0])
Here. I did a quick "15 minute" look. May not be accurate
That gives us $200-$280 before counting the power supply and buck converter.[0] When I wrote the code only me and god understood what was going on. But as time marched on, now only god knows.