RIEGL Deep Learning Suite
RIEGL Deep Learning Suite is a set of AI plug-in tools for RiSCAN PRO and RiPROCESS. It blurs faces and licence plates in survey imagery and classifies airborne and mobile LiDAR point clouds automatically.
What is RIEGL Deep Learning Suite?
RIEGL Deep Learning Suite is not a standalone application. It is a family of artificial intelligence tools that RIEGL, the Austrian LiDAR manufacturer, added to its own processing software in February 2024, alongside RiSCAN PRO 2.19 and RiPROCESS 1.9.4. The vendor's own software catalogue lists it under the label plug-in, and everything it does happens inside one of those two host programs: RiSCAN PRO for terrestrial scanning, RiPROCESS for kinematic data captured from aircraft, drones and vehicles.
The suite currently covers two jobs that RIEGL says appeared most often on customer request lists. The first is image anonymisation. The RIEGL Image Anonymizer finds faces and licence plates in the photographs attached to a scan project and renders them unrecognisable through adjustable pixelation. It can be triggered at any point in the chain — before colouring the point cloud, or before publishing a RiPANO project — and it runs across several records, scan positions and cameras in one pass. The driver is regulatory: survey imagery frequently has to be anonymised before it leaves the service provider, and certainly before it is made public.
The second job is point cloud classification. For airborne data, the suite assigns ASPRS classes for ground, vegetation and buildings across urban or mountainous areas. For mobile data, it goes further and separates road, pavement, cars, poles, traffic signs and lights and road markings, with a dedicated highway model that adds gantries, walls and tunnels. RIEGL attributes the performance to Sparse CNN, a convolutional architecture designed for very large point clouds, accelerated by CUDA graphics cards. Models are pre-trained, so parameter setting is described as minimal. A fourth tool, in RiSCAN PRO, uses a deep learning module to recognise rectangular pattern targets in a terrestrial cloud.
Processing is strictly local: the vendor states the anonymiser runs on your own device. Access is closed, though. The software downloads sit behind the customers-only Members Area, RIEGL packages ship exclusively with RIEGL scanners, and the classification tools need a separate licence obtained from a reseller.
What it does
- Detect and blur faces and licence plates automatically in the imagery of a scan project
- Classify airborne LiDAR point clouds into ground, vegetation and building classes
- Classify urban mobile LiDAR into road, pavement, car, pole, traffic sign, traffic light and road markings
- Apply a dedicated highway classifier that adds gantry, wall and tunnel classes
- Identify rectangular pattern targets in terrestrial point clouds and fit them to extract their centre point
- Run the blurring pass in batch across several records, scan positions and cameras
- Adjust the shape and style of the pixelation applied to detected objects
When to use RIEGL Deep Learning Suite / When not to
A quick filter to help you decide if RIEGL Deep Learning Suite is the right fit.
When to use RIEGL Deep Learning Suite
- Existing RiPROCESS users processing airborne, drone-borne or mobile LiDAR campaigns
- Existing RiSCAN PRO users working with terrestrial scan projects
- Mobile mapping contractors who must remove faces and licence plates before delivering or publishing imagery
- Wide-area mapping teams that need ground, vegetation and building classes without manual editing
- Road and highway survey crews extracting street furniture, markings, gantries, walls and tunnels
When not to use RIEGL Deep Learning Suite
- Owners of non-RIEGL scanners, since the software packages are bundled exclusively with RIEGL hardware
- Buyers who need a published price or an online purchase, as every module is quoted through a reseller
- Teams looking for a cloud service or a public API, because everything runs on a local Windows workstation
- Workstations without a CUDA-capable graphics card, which the vendor ties to the announced performance
- Anyone hoping to evaluate the product first, as no free plan and no free trial are offered
How to use RIEGL Deep Learning Suite
A typical end-to-end flow, from setup to results.
- Acquire a RIEGL scanner or scanning system, since the software packages ship only with RIEGL hardware
- Ask your local reseller, or sales@riegl.com for direct purchases, for a licence covering the module you need
- Choose the licence type: a floating licence tied to a USB dongle, or one tied to a scanner serial number
- Install RiSCAN PRO for terrestrial work, or RiPROCESS for airborne, drone and mobile work, from the Members Area
- Import the licence key into RIEGL LicenseCenter, or into the LicenseServer Administration Center for floating licences
- Check that the workstation has a CUDA-capable graphics card before running classification
- Open the scan project and load the raw LiDAR data and imagery
- Run the Image Anonymizer on the selected records, scan positions and cameras, choosing the pixelation style
- Run the airborne or mobile classifier with its pre-trained model and review the resulting classes
- Export the classified cloud to LAS, LAZ, E57 or COPC for the rest of your workflow
Pros & Cons
Pros
- Runs inside the RIEGL software the team already uses, with no parallel toolchain to set up
- Processing is entirely local, so scan data never leaves the workstation
- Pre-trained models keep parameter setting to a minimum, according to the vendor
- Answers a concrete regulatory obligation by anonymising personal identifiers in survey imagery
- Separate classifiers tuned for airborne, urban mobile and highway datasets rather than one generic model
- Sparse CNN and CUDA acceleration are aimed specifically at very large point clouds
- Backed by a manufacturer with more than forty-five years in laser scanning and a TÜV-certified ISO 9001 quality system
Cons
- No price is published anywhere; every module has to be quoted through a reseller
- Tied exclusively to RIEGL hardware, so it cannot be used with another manufacturer's scanner
- Classification needs an additional licence on top of the host software licence
- Neither a free plan nor a free trial is offered
- Product documentation and downloads sit behind a customers-only area, so evaluation before purchase is impossible
- A CUDA graphics card is required for the performance the vendor describes
- No terms of use or licence conditions are published, leaving data ownership and reuse undocumented
Pricing & Plans
There is no free plan and no published price. RIEGL states that its software packages are supplied exclusively with the corresponding RIEGL laser scanners and scanning systems, and that the AI airborne classification tool requires additional licensing to be arranged with a local reseller. Direct customers are directed to sales@riegl.com. Consequently no entry price, currency or billing unit can be reported for this product: the amount depends on the modules selected, the licence type and the maintenance period agreed in the quotation.
- DLS Image Content Anonymization — licensed module of the RIEGL Deep Learning Suite
- price on quotation
- DLS Point Cloud Classification ALS — licensed module for airborne data
- price on quotation
- DLS Point Cloud Classification MLS — licensed module for mobile data
- price on quotation
- Single Instance Floating License (SIL) — bound to a USB dongle and managed by RIEGL LicenseServer across a local network
- Scanner Serial Number License (SNL) — bound to one scanner serial number and usable in parallel on several workstations
- Light License — reduced-price variant limited to data from miniVUX-1UAV
- miniVUX-1DL
- miniVUX-2UAV
- miniVUX-3UAV and VZ-i series scanners
- Software maintenance period — sold with every licence
- once it expires the licence keeps working but newer versions are no longer available
Data, GDPR & hosting
A consolidated view of how RIEGL Deep Learning Suite handles your data.
GDPR overview
The site publishes an Article 13 privacy notice, last updated on 3 September 2026 and served through a third-party compliance widget. The controller is RIEGL Deutschland Vertriebsgesellschaft mbH in Gilching, Germany, with Erwin Burth named as privacy coordinator; no Article 37 data protection officer is appointed. Data subject rights under Articles 15 to 21, 7(3) and 77 are set out, exercised by writing to the address given, and the competent authority is named. Two limits matter. The notice covers website visits and enquiries only — never the data processed by the software — so no GDPR fact drawn from it applies to your scan projects. And RIEGL nowhere claims to be GDPR compliant in so many words. The product itself is presented as a means of meeting the regulation, since its Image Anonymizer exists to make personal identifiers unrecognisable.
Who owns the data?
No terms and conditions, licence agreement or end-user contract is published anywhere on riegl.com, so ownership of the data you process is not documented publicly. Five candidate paths were probed and all return a genuine 404, and the fifteen-page sitemap lists none. The only privacy policy on the site covers website visitors and business enquiries, not the data handled by the software. What the vendor does state, in its own release note, is that the Image Anonymizer operates locally on your device so that sensitive material stays with you. Scan projects therefore remain on your workstation, and nothing describes a transfer of customer data to RIEGL. Ownership terms should be requested from the reseller before contracting.
Reuse rights
Because no licence terms are published, nothing on the site restricts or authorises what you may do with the classified point clouds and anonymised images the suite produces. In practice the outputs stay inside your own project: classified clouds can be exported to standard formats such as LAS, LAZ, E57, COPC, SHP and ASCII for downstream use in third-party GIS, CAD and BIM software, and anonymised imagery can be coloured onto the cloud or published through RiPANO. No approval step involving RIEGL is described at any point. Since the contractual position is simply absent from the public site, reuse rights should be confirmed in the quotation or licence certificate rather than assumed.
Data retention & training
Hosting summary
There is no hosting to describe on the product side. The suite runs inside desktop software on the customer's own Windows workstation, and RIEGL states explicitly that the Image Anonymizer operates locally on your device. No cloud service, public API or application subdomain exists, so no country or region of data hosting is declared for customer data, and none could be. The marketing website is a separate matter. It is operated by RIEGL Deutschland Vertriebsgesellschaft mbH in Germany; the domain's address record points to a German host, while the pages themselves are delivered through a content delivery network. The website privacy notice flags possible transfers to the United States through Google Analytics. All of that concerns visitors to riegl.com, not the scan data processed by the software, and should not be read as a statement about the product.
Things to keep in mind
Risks and trade-offs to weigh before adopting RIEGL Deep Learning Suite.
- The company named in the site's legal notice is RIEGL's German distribution subsidiary in Gilching, while the product's own release note is signed by RIEGL Laser Measurement Systems GmbH in Horn, Austria, which is also the domain registrant: the identity fields describe the site operator, not the software author
- AI anonymisation is not infallible; a face or plate the model misses is still personal data once the dataset is delivered or published, and the legal responsibility stays with the data controller
- Automatic classification produces plausible but unguaranteed classes, and no accuracy metric is published, so human review remains necessary before any deliverable leaves the office
- The suite locks the workflow to RIEGL hardware and RIEGL host software, which deepens an existing dependency rather than reducing it
- With no published price for the module or for maintenance, no comparison with alternatives is possible before requesting a quotation
- Once the maintenance period ends, no further versions can be installed, so the real cost is recurring even though it is never displayed
- No terms of use or licence conditions are published, so data ownership and reuse rules cannot be checked before signing
Setup & Integrations
Technical difficulty
Moderate, and the difficulty lies in licensing rather than in the AI. You must already run RiSCAN PRO or RiPROCESS, then install a licence: either a USB dongle with RIEGL LicenseServer administered as a Windows service on the network, or a key tied to a scanner serial number imported on each workstation. A CUDA-capable graphics card is needed for the announced performance. Once that is done the tools themselves are undemanding, since the models are pre-trained and RIEGL says parameter setting is reduced to a minimum. Initial training is included with any RIEGL purchase.
Deployment
Behind RIEGL Deep Learning Suite
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
Is RIEGL Deep Learning Suite a standalone program?
What exactly does it do?
Which classes does the classification produce?
Does my data go to the cloud?
What hardware do I need?
How much does it cost?
Can I use it with a scanner from another manufacturer?
Is there a free trial or a free plan?
How do I get a demonstration?
How does the licensing work?
Should you pick RIEGL Deep Learning Suite?
RIEGL Deep Learning Suite is a narrow, well-targeted product rather than a general purpose AI platform, and it should be judged on that basis. It exists to remove two recurring chores from an existing RIEGL workflow: making personal identifiers unrecognisable in survey imagery, and turning raw point clouds into classified ones. Both are jobs the vendor says customers asked for repeatedly, and both are delivered where the work already happens, inside RiSCAN PRO and RiPROCESS.
The technical case is credible on paper — Sparse CNN, CUDA acceleration, pre-trained models, separate classifiers for airborne, urban and highway data — but RIEGL publishes no accuracy figures, no benchmark and no independent evaluation. A surveyor should therefore treat the classes as a strong first pass to be checked, not as a finished deliverable, and the same caution applies to anonymisation: a missed face or plate is still published personal data, and the responsibility stays with whoever ships the dataset.
The commercial picture is the real filter. The suite is inaccessible unless you already own RIEGL hardware and the matching host software, the classification tools need a further licence, there is no free plan and no trial, and not a single amount is published. Documentation sits behind a customers-only area, so there is no way to evaluate before contracting. Buyers should also note that the site is operated by RIEGL's German distribution subsidiary while the product itself is authored by the Austrian parent in Horn, and that no terms of use or licence conditions are published at all.
For an existing RIEGL customer, this is a logical and probably worthwhile extension of a chain they already run. For anyone else, it is not a product that can be bought, tested or even properly compared.
- Choosing a selection results in a full page refresh.
- Opens in a new window.