[DoofusOfDeath]

Anyone have a LIDAR (or similar) product suggestion?

I'm getting ready to start a hobby project that involves scanning the interior surfaces of a house. Ideally the accuracy would be at least 1/16" (1.5mm), including any scan-stitching required because the sensor had to be moved around.

I've seen a few promising products, but none stands out as a perfect match.

[danpat]

I've had some success with some of the off-the-shelf laser distance measurement modules, like the ones Bosch sell at big-box hardware stores:

https://www.boschtools.com/us/en/boschtools-ocs/laser-measur...

They have 1-2mm accuracy and quite good precision, particularly if your environment is stable (temperature, movement, lighting). These modules use an interferometry approach rather than time of flight to achieve their accuracy.

Several of them have a Bluetooth interface which you could reverse-engineer. The work then would be creating a turret to rotate the unit around a known centerpoint and take a bunch of samples. It'd be slow, but it works.

There are also a bunch of modules available on Alibaba for a few tens of dollars that have serial interfaces and seem to have similar performance - they often have 10s of Hz sample rates, so you could speed up scanning quite a lot. They exhibit similar accuracy to the boxed units I've bought, but require you to be comfortable with things like SPI and soldering.

[jefft255]

1.5 mm is really ambitious. This is close to metrology-level precision. I wouldn't look into sensors aimed at robotics/autonomous vehicles. Try Creaform, Riegl stuff, etc... Also what you're trying is far from new so maybe get acquainted with what other businesses are doing (e.g. https://www.bentley.com/en/products/brands/contextcapture).

[paol]

If you just want the result of the scan (not clear if the hobby project is the scanning itself or something else) you should be able to contract it as a service, or just rent the scanner. This is common for architectural scan work.

[lnsru]

1,5 mm accuracy isn’t doable that easily. There is company called Photoneo having structured light system with such advertised accuracy. But the scan volume is rather limited.

Working on time-of-flight Stereo camera in my spare time. Few centimeter error is very normal. It was shocking at the beginning, but I now understand why bin picking is still hard task.

[briffle]

A project I wanted to play with 10 years ago, and didn't have time or money, was a tool you could set in a room, and it would put laser 'dots' along the ceiling where crews should hang the parts for a drop ceiling, to minimize cuts of both hanger equipment, and the ceiling tiles.

[DoofusOfDeath]

That's closely related to what I'm going for. Ultimately I'm looking to make an A.R. system that guides various house remodeling tasks, including framing and floor-leveling.

[maxbaines]

https://store.intelrealsense.com/buy-intel-realsense-lidar-c...

[Symmetry]

You'll need to do a bunch of averaging to get that sort of precision out of one of those, they tend to have an error proportional to distance that's generally a lot higher than the OP's 1.5mm target. But I agree that that's the right starting place.

[mikeha]

Take a look at Leica RTC 360. It's a survey grade terrestrial lidar scanner.

Angular accuracy 18” Range accuracy 1.0 mm + 10 ppm 3D point accuracy 1.9 mm @ 10 m 2.9 mm @ 20 m 5.3 mm @ 40 m

(disclaimer: I work for Leica)

[baybal2]

mm wave radar + sar postprocessing? Given you have a 100% static scene, it should work great

[DoofusOfDeath]

Are you aware of any low-cost, consumer-oriented products for that?

Perhaps my Google-Fu is weak today, but I'm only finding research / military projects.

[Reelin]

Is something like this (http://www.ti.com/product/IWR6843) relevant? There are also older products such as this (https://www.digikey.com/en/product-highlight/a/acconeer-ab/a...).