National mapping
A whole season, classified to your specification
We fit a classifier to your programme specification and run the season on it: your class catalog, your tile scheme, your delivery windows, with the quality control taken on where you want it off your own desk.
What we do to it
Your catalog, your specification, at programme scale
We have classified ALS for the companies that fly it for years, and the decisions in a workflow come from that — where a rule beats a model, how a block gets sampled for training, what the post processing has to catch. We work from a modular training base, trained on datasets from Europe, the USA, New Zealand and Australia, so fitting a classifier to your class list is measured in hours.
On the Digital Twin Germany programme the line runs at over 5,000 km² in 24 hours, which on that data is up to six terabytes.
Your catalog, your sensors, tuned on the existing base.
The platform starts more machines when jobs queue, so a large block is a scheduling question.
Quality control against your acceptance criteria, managed by us where you want it off your own desk.
Objects and key points where the programme needs geometry as well as classes.
On real data
Water, coast and everything on it
One catalog has to hold over open water, a railway, a village and the gardens behind it, and hold the same way two blocks later.
A dense estate at full density
Buildings, vegetation, ground and vehicles separated across a housing estate. A national catalog proves itself on a scene like this one.
The objects a programme is actually after
High voltage lines, wind turbines, photovoltaic systems, towers and pipelines. Each one is a shape the model has to find in a scene full of things that look like it.
Deliverables
What the survey is for
Every point in the cloud carries a class. That is what the filtering and the cleaning run on, and what every product under it is built from.
DTM, DSM and the normalised surface model as rasters on your tile scheme. Most of a programme specification is written in those three.
Canopy height across the block, out of the classified cloud, for the forestry and biomass parts of the programme.
Buildings and vegetation compared class by class against the previous campaign, which is the reason a programme reflies at all.
What lands in your hands
The output
- Airplane
- Helicopter
- Drone
airborne lidar, at programme scale
- LAS
- LAZ
- DXF
- Shapefile
- CSV
- GeoJSON
classified cloud and objects
The class catalog
Detect various objects on airborne laser scans. The model was optimized for point clouds with intensity values and average densities of 10 to 50 ppm². If no CRS is specified, the point cloud unit must be meter.
10 classes
recommended
- Intensity
required in the file
IDs follow the ASPRS numbering in the LAS specification. Classes above 18 are ones we added for this catalog.
| ID | Class |
|---|---|
| 2 | Ground |
| 5 | Vegetation |
| 6 | Building |
| 12 | Power Pole |
| 14 | Power Lines |
| 15 | Transmission Tower |
| 18 | Noise |
| 19 | Street Furniture |
| 20 | Vehicles |
| 22 | Other |
Where it runs
On Microsoft Azure inside the EU by default.
A separate tenant in the Azure location your rules require.
The whole stack on your own hardware, air gapped where that is the requirement.
Beyond the standard catalog
Custom model training and finetuning for your dataset
We train on your data and tune the model to the specification you work to. Classes get added, split or cut. A photogrammetric cloud gets a model of its own.
A project like yours
What the client said
“Together with BSF Swissphoto, Pointly worked in a focused manner toward the project goal of achieving a classification accuracy of over 95 %, and developed a tailored workflow that was precisely aligned with the project requirements and target structures.”
Questions
Before you commit a survey
- How much can you sustain in a day?
On the Digital Twin Germany programme the line runs at over 5,000 km² in 24 hours, on airborne data at roughly 40 points per m². That is up to six terabytes a day. A larger block runs on more machines and finishes in the same window.
- Can you work to our class catalog and our tile scheme?
Yes. We fit the classifier to your programme specification, your tile scheme and your delivery windows. We adapt the existing training base to a new catalog in hours.
- Where does our data sit while it runs?
Your data sits on Microsoft Azure inside the EU by default, or in a separate tenant in the region your rules require. Where it has to stay on your own network, the same stack runs there as the Pointly Box.
See it on your own data
Send a sample and we will run the classifier on it, so you can judge the result rather than the description.