¶ autonomous drone & end-to-end automatic forestry inventory system
| Company | Deep Forestry |
| Country | Sweden |
| Website | https://www.deepforestry.com/ |
| Category | Planning & Inventory › Remote sensing & mapping |
¶ Overview
An autonomous drone-based end-to-end automatic forestry inventory system integrates unmanned aerial vehicles (UAVs) with advanced sensors, artificial intelligence, and cloud-based data processing to deliver comprehensive, automated forest resource assessments without requiring manual ground-truthing at scale. These systems typically combine RGB cameras, multispectral sensors, and LiDAR payloads mounted on fixed-wing or multi-rotor drones to capture high-resolution aerial data across large forested areas in a single flight or coordinated multi-drone mission. Onboard or edge-computing modules perform initial data filtering, while back-end machine learning pipelines — often using deep convolutional neural networks or point-cloud processing algorithms — automatically segment individual tree crowns, estimate stand height and basal area, classify species composition, and detect stress, disease, or pest damage. The "end-to-end" designation reflects the seamless pipeline from autonomous flight planning and data capture through to the delivery of structured inventory outputs such as georeferenced tree maps, volume estimates, and growth projections, all without manual photointerpretation steps. In Australian forestry, organisations such as the Forest and Wood Products Australia (FWPA) and state-level agencies have trialled UAV inventory methods across plantation softwoods in Victoria and hardwood estates in Western Australia, finding drone-derived estimates comparable in accuracy to traditional plot-based cruising at a fraction of the cost and time. Key challenges include regulatory compliance under CASA's remotely piloted aircraft rules, battery endurance over large compartments, and achieving sufficient point-cloud density beneath dense canopy. Emerging developments include the use of AI-guided adaptive sampling, where the drone autonomously identifies and revisits anomalous zones flagged during the initial pass, and integration with forest management information systems to automatically update standing timber inventories in near-real-time, supporting more responsive harvest scheduling and carbon accounting.
¶ Assessment
A highly suitable tool that offers major gains in speed and detail of forest inventory and early issue detection, though its value depends on adoption feasibility and operational integration.
¶ Details
| Field | Value |
|---|---|
| Company | Deep Forestry |
| Country | Sweden |
| Type | Company |
| FWPA RD&E | 7.4 |