🔭 OpenSLAM
DIY Architectural & Surveying LiDAR SLAM Scanner
Compact open-source 3D scanner for architectural and surveying work built with Livox Mid-360 and NanoPi Zero 2.
Scan any space and get a full 3D point cloud — built with the free in-house A2 mapper,
with Share Studio kept as a fallback.
📣 Questions & support → Telegram @a2blog
📋 Table of Contents
🆕 What’s New
- Own SLAM engine is released: A2 mapper. The point cloud is now built in-house, with no third-party software in the chain. The free A2 mapper for Windows and Android reads the scanner’s
.bin recordings directly and writes a .las point cloud plus a trajectory. It can also pull the trajectory onto surveyed control points (GCP / RTK) and output the cloud in their coordinates. Share Studio stays supported as a fallback route (see Software & Setup).
- Storage: eMMC → microSD. The board no longer needs a special eMMC-equipped NanoPi. Any regular NanoPi Zero 2 works — scans are written to a plain microSD card (64GB+, U3/V30 or faster). Cheaper board, easier to source, and the card itself is now a swappable, off-the-shelf part.
- App rewritten. The Monitor screen now draws a live point cloud + trajectory in plan view while you walk, shows scanning-quality metrics, and — the thing everyone actually wants to know in the field — how much recording time is left, in hours and minutes, based on free space. Files got multi-select copy/delete. Settings got lidar model switching, setting the board’s clock from your phone, and updating the scanner’s onboard software right from the app (see below).
- Converter takes a whole folder now.
A2 Bin to ShareStudio.exe accepts multiple .bin files at once, and can now trim both the start and the end of a recording (previously only the start).
- Remote setup got its own tool. If you go with Option A (contact us for setup), we now flash and configure your SD card in a remote session with a dedicated installer — faster and more consistent than manual steps over chat.
- Panorama integration is done. Insta360 camera support is complete and working well — point clouds get their real color, and full panoramic tours build cleanly. Demo: tour.a2-lab.ru/dtyevqr7nd7a. A separate repo update is coming for Insta360-based E57 panorama generation, in addition to point-cloud coloring — see Panorama Color & Tours.
Older hardware (eMMC board) and the previous app/converter are not deleted — see Previous Version.
💡 What is this?
A handheld SLAM scanner designed for architects, surveyors, and construction professionals — capture spatial datasets and build point clouds from them with the free A2 mapper. Share Studio remains available as a fallback processing route.
The scanner is open hardware: source the components, 3D-print the parts, and assemble it yourself. For software, you can either handle the dataset conversion workflow on your own or contact us for setup and support.
📦 Components
| Part |
Notes |
| Livox Mid-360 |
Main LiDAR sensor. Mid-360S also supported |
| NanoPi Zero 2 |
Wi-Fi module required. No eMMC needed anymore — see storage below |
| microSD card |
64GB minimum, U3/V30 (or faster) speed class. This is the scanner’s storage now — regular eMMC boards are no longer required |
| Powerbank |
2 outputs required: one 10W+, one with 12V or 20V PD output. Tested: HPS99 ✅ · HPS36 should work (lighter, untested) |
| LiDAR cable + USB-C PD trigger |
Rigid 0.5m (durable) or flexible 0.2m (compact). USB-C PD trigger 12V or 20V — soldered to power end |
| Part |
Notes |
| microSD card (V30) |
See storage requirement above — this is the card itself |
| USB-C PD trigger board |
Wired into the power cable, sets the powerbank’s output voltage |
| V-mount plate (optional) |
Only needed if you’re mounting a panorama camera on top — see Panorama Color & Tours |
| M3×8 × 4pcs |
LiDAR to body |
| M2.5×12 × 4pcs |
NanoPi to body |
| Fixator (1/4" screw) |
Handle / tripod mount |
💬 Not sure which components to pick? Ask in our chat — t.me/a2blogchat
📐 Need a custom form factor with this same architecture? We can design it — write us
🖨️ 3D Printed Parts
Material: PLA or PETG · Layer height: 0.2mm · Supports: Body only
The head was redesigned for the new build — if you printed the old one for an eMMC-era board, it still works, but grab the new head.stl for new builds.
🔧 Assembly
Only a screwdriver with bits for M3 and M2.5 screws needed. Assembly takes 5–10 minutes.
▶️ Watch assembly on YouTube Shorts
💻 Software & Setup
The repository includes ready-to-use software in the /software folder:
| File |
Description |
A2 mapper — Windows .exe and Android .apk (separate repository) |
Recommended. Builds the point cloud (.las) and the trajectory straight from the scanner’s .bin recordings, with optional control-point (RTK / GCP) correction. Free to use under its own freeware license — see the A2-slam repository. |
| Share Studio v2.2.4 (fallback · not included, third-party) |
Point cloud processing software by SHARE3DCAM — an alternative way to generate the final point cloud from the converted scan. The converter is built for v2.2.4; newer versions (2.6.x) are not compatible yet. Download v2.2.4 from the official SHARE website: Support → Software → SHARE PointClouds Studio → Version History. Having trouble? Message me on Telegram @Al_And_Leon. |
AtlasScanner.apk |
Android app — connect to the scanner, watch the live map and metrics, manage files, update the scanner’s software |
A2 Bin to ShareStudio.exe |
Fallback route: converts scan output to Share Studio format. Takes multiple .bin files at once, trims start and end |
A2 SD Card Installer.exe |
Flashes and configures a microSD card — used for remote setup (Option A below) |
App screens
Monitor — before connecting, after connecting, and during a scan. While scanning it draws a live plan-view point cloud with your walked trajectory, four quality metrics (angular rate, vibration, point quality, return rate), and how much recording time is left, in hours and minutes, based on free space on the card.
Files — the scan list with free space at a glance, multi-select Copy (to USB or straight to your phone) and Delete with a confirmation.
Settings — pick the lidar model config (Mid-360 / Mid-360S), set the board’s clock from your phone, and update the scanner’s onboard software right from the app. New scanner updates will be delivered this way going forward — no more separate tools or manual flashing for software-only changes.
Converter & remote installer
The converter now accepts a batch of .bin files in one go, and can trim both the start and the end of a recording. The installer is what we use to remotely flash and configure your microSD card (Option A).
We flash the card, configure everything, and the software above works out of the box. Setup now happens in a short remote session using the SD Card Installer tool above — quick and consistent.
→ Write us on Telegram @a2blog
Option B — DIY setup
If you want to configure everything yourself, here’s what needs to be done on the NanoPi Zero 2:
- Install ROS
- Install Livox SDK
- Install Livox ROS Driver 2
- Write a script for automatic scan start/stop
- Configure and broadcast a Wi-Fi access point on the board
- Develop a smartphone app to control start/stop over Wi-Fi
- Process the recordings: A2 mapper reads the A2
.bin format and ROS1 bags with the /livox/lidar and /livox/imu topics, or you can develop your own converter to a Share Studio-compatible format
This is a non-trivial setup. If you’re not sure — just ask us, it’s faster.
You can prepare your own recordings yourself using ROS — the list above is only an outline. To process them with A2 mapper, see the recording requirements in that repository.
📡 How Scanning Works
▶️ Full process — from power-on to point cloud (YouTube Shorts)
The section below applies to the our setup version (APK + converter). If you configured everything yourself — you have your own pipeline.
Starting a scan
- Connect the powerbank — scanning starts automatically
- Wait 5 seconds with the scanner still and flat before picking it up
- Slowly lift it and start walking through the space
- To stop: use the app or simply disconnect the power
💡 For the best result with A2 mapper: stand still for about 10 seconds at the start, walk the route, come back to the exact spot you started from, and stand still for about 10 seconds again before you stop. The start and the finish are then tied together and the loop is closed.
Connecting the app
- Disable VPN and mobile data on your phone
- Connect to Wi-Fi network A2-scanner, password
12345678
- Open the app — once the LiDAR is powered on you’ll see a live map of the space, quality metrics, and recording time left
- The app also shows scanning recommendations during capture
▶️ App interface and scanning tips (YouTube Shorts) (recorded on an earlier app version — the flow is the same, the screens now show more)
Getting your data
- Copy the scan straight from the app — to a USB drive plugged into the NanoPi, or directly to your phone
- Open the
.bin file(s) in A2 mapper and press Build — you get a .las point cloud and a trajectory (the repository explains the optional control-point correction)
- Fallback: if you prefer Share Studio, run
A2 Bin to ShareStudio.exe on your PC to convert the file(s) — a whole folder at once if you like — and import them into Share Studio to build the point cloud there
Accuracy vs total station
▶️ SLAM scanner vs total station — measurement accuracy comparison (YouTube)
Independent comparison of OpenSLAM point cloud data against total station survey measurements — useful if you’re evaluating the scanner for surveying applications.
🎨 Panorama Color & Tours
Insta360 camera integration is complete — mount a panorama camera on the optional V-mount plate, and the pipeline colors the point cloud from the 360° footage and builds a full panoramic tour.
▶️ Live demo — colorized point cloud + panoramic tour
A separate repository update is coming for Insta360-based E57 panorama generation, in addition to the point-cloud coloring shown above.
🗄️ Previous Version
The eMMC-based hardware revision and the older app/converter are kept, not deleted — they’re tagged and archived as a GitHub Release so you can still reach them if you need to roll back or you’re maintaining an older unit:
→ Legacy release: eMMC board, old app & converter
Everything on this page (main branch) is the current, actively maintained version.
📁 Repository Structure
OpenSLAM/
├── README.md
├── stl/
│ ├── head.stl
│ ├── guard.stl
│ ├── cover.stl
│ └── stand.stl
├── software/
│ ├── AtlasScanner.apk
│ ├── A2 Bin to ShareStudio.exe
│ └── A2 SD Card Installer.exe
└── images/
├── hero.png
├── components-main.jpg
├── components-small.jpg
├── components-printed.jpg
├── app-monitor.jpg
├── app-files.jpg
├── app-settings.jpg
├── converter-installer.jpg
├── converter-detail.jpg
└── sd-installer-detail.jpg
🛣️ Roadmap
This is not a finished project — it’s an actively evolving platform.
Point clouds are colored, panorama tours work end to end, and the in-house SLAM engine — A2 mapper — is released. Here is where everything stands:
| Feature |
Status |
What it gives you |
| GNSS integration |
✅ Implemented |
Geo-referenced point clouds placed directly into world coordinate systems |
| Ground control points (GCP) |
✅ Implemented |
Mark surveyed reference points during a scan; A2 mapper corrects the drift between them |
| Own SLAM engine |
✅ Implemented |
A2 mapper — an end-to-end in-house pipeline from the scanner’s .bin to the point cloud. Share Studio stays as a fallback |
| 3D Gaussian Splatting (3DGS) |
🔄 In progress |
Photorealistic, navigable scenes built from scan data |
| Web service for cloud uploads |
📅 Planned |
Upload point clouds — including colored, panorama-tagged ones — straight from the field |
📣 Follow updates in our Telegram — t.me/a2blog
📄 License
MIT License — free to use, modify, and share. See LICENSE.
The software built on the scanner’s data, A2 mapper, is distributed separately under its own freeware license.
Telegram @a2blog — questions, setup requests, feedback.
Commercial inquiries (board setup, scanner assembly, custom builds) — [email protected]
If this helped you, give the repo a ⭐
OpenSLAM is an independent project and is not affiliated with SHARE3DCAM. “SHARE” and “PointClouds Studio” are trademarks of their respective owners.
评论