Flai
Flai is an Artificial Intelligence (AI) solution intended for automatic LiDAR point cloud classification. Its main goal is to automate the extraction of custom,...
Last verified:
What is Flai?
Flai is an AI-powered platform for automatic classification of LiDAR point clouds. It transforms unstructured large LiDAR data sets into actionable insights with minimum manual effort, enabling users to perform semantic segmentation of point cloud data quickly and accurately.
Key features include four pre-trained AI models (Aerial Mapping, Forestry, Mobile Mapping, and Indoor Stockpile) with 40+ different semantic classes, advanced data processing with over 40 modular processors for comprehensive point cloud manipulation, custom classifier training for extracting tailor-made features, and flexible deployment options from cloud-based SaaS processing to self-hosted solutions. The platform supports UAV surveys, aerial and mobile surveys, and works with various sensors including airborne, terrestrial, mobile, and UAV LiDAR systems.
Flai is designed for GIS consultants, geospatial professionals, forestry experts, UAV survey operators, urban planners, power network maintenance teams, and larger companies working with LiDAR mapping of forest areas, powerlines, cities, open mines, and natural disaster damage mapping.
The platform offers manual annotation tools directly on the cloud-based machine learning platform and annotation-as-a-service for data labelling to improve AI model quality and applicability.
Flai pricing
Pricing model: Freemium
Flai uses transparent pay-per-use pricing based on volume of data and complexity of processing flow. Users create a free account and process for free, only paying when downloading results. Flai offers multiple pricing tiers: Free trial (25GB storage), Starter (25GB), Growth (100GB), and Professional (1000GB). Plans start from €20/month. Batch processing is project-based for mid-to-large projects with price defined per square kilometer/mile or line kilometer/mile based on project size, point density, and number of classified categories. Self-Hosted option requires annual license with one-time payment. Custom AI model training pricing is defined by project requirements. VAT not included. Pricing calculator available on website for independent cost estimation.
Flai pros
- AI-powered automatic point cloud classification with minimal manual effort
- Four pre-trained AI models covering aerial, forestry, mobile, and indoor mapping
- 40+ different semantic classes for comprehensive classification
- Over 40 modular processors for advanced data processing
- Custom classifier training for extracting custom features
- Flexible deployment: cloud-based SaaS or self-hosted solution
- Fast processing - cuts manual classification time by factor of four
- Supports UAV, aerial, and mobile LiDAR surveys
- Works with airborne, terrestrial, mobile, and UAV sensors
- Dynamic classification lets you select desired categories
- 3D point cloud visualization directly in web application
- Manual annotation tools built into the platform
- Annotation-as-a-service available for data labelling
- Batch processing via AWS S3 or FTP bypassing upload/download
- 40% faster processing time with batch processing approach
- Pricing calculator for estimating project costs independently
- Free tier available to try before paying
- Trusted by State of Neuchâtel Switzerland and other organizations
- Integration seamlessly into existing GIS workflows
Flai cons
- Pay only when downloading results - processing costs may accumulate
- Self-hosted requires Ubuntu 18.x or greater operating system
- Self-hosted needs minimum 8 CPU cores, optimally 16 cores
- Self-hosted requires minimum 16 GB RAM, optimally 64 GB
- Self-hosted needs Nvidia GPU with minimum 11 GB vRAM
- Pricing based on data volume and processing complexity - not fixed
- Free trial storage limited to 25GB
- Some data restrictions apply for SaaS depending on geography
- Custom AI model training requires contacting support for pricing
- Batch processing S3/FTP requires correct EPSG code specification
Frequently asked questions about Flai
What is Flai and what does it do?
Flai is an AI-powered platform for automatic classification of LiDAR point clouds. It transforms unstructured large LiDAR data sets into actionable insights with minimum manual effort by performing semantic segmentation (classification) of LiDAR points into desired categories using pre-trained AI models.
What AI models does Flai offer?
Flai offers four distinct pre-trained classification models: Aerial Mapping AI model (for ground, vegetation, buildings, wires, powerline towers, bridges, vehicles, fences, etc.), Forestry AI Model (for ground, vegetation, tree trunks, fallen trees), Mobile Mapping AI Model (for roads, sidewalks, traffic islands, buildings, trees, traffic lights, pedestrians, vehicles), and Indoor Stockpile AI Model (for stockpile and other man-made objects).
How many semantic classes does Flai support?
Flai offers 40+ different semantic classes across its four pre-trained AI models, including 19 categories for general mapping, 6 for forestry, and 21 for mobile mapping.
Can I train my own custom classifier?
Yes, Flai allows you to train tailor-made custom classifiers to extract custom features from point cloud data. You can retrain AI models for specific objects or features relevant to your use case, and annotation-as-a-service is available for data labelling to improve model quality.
What deployment options does Flai offer?
Flai offers flexible deployment options including cloud-based Software as a Service (SaaS) through the Flai web platform and Self-Hosted solution for running Flai algorithms locally on-premises for enhanced data security within your organization.
How is Flai pricing structured?
Flai uses pay-per-use pricing based on volume of data and complexity of processing flow. Users process for free and only pay when downloading results. Pricing can be per square kilometer/mile for batch processing, with plans starting from €20/month. A transparent pricing calculator is available for estimating project costs.
What storage limits are available?
Storage limits vary by plan: Free trial (25GB), Starter (25GB), Growth (100GB), and Professional (1000GB).
What sensors does Flai support?
Flai supports LiDAR data from airborne, terrestrial, mobile, and UAV sensors for aerial mapping, forestry, mobile mapping, and indoor stockpile classification applications.
How fast is Flai processing compared to manual classification?
Using Flai cuts down the time for manual point cloud classification by a factor of four, radically improving turn-around and project profitability. Batch processing achieves approximately 40% decrease in processing time by bypassing upload, download, and visualization on the web app.
How do I get a self-hosted trial license?
Write to [email protected] to get a 10-day trial period for the self-hosted solution. You can utilize the Flai self-hosted solution without any restrictions during the trial period.