¶ Visual positiong
| Company | Immersal |
| Country | Finland |
| Website | https://immersal.com/techModellogy |
¶ Overview
Visual positioning is a localisation technique that determines the position and orientation of a machine or device by analysing imagery from one or more cameras, rather than relying solely on GNSS satellite signals. In forestry and logging applications, where dense canopy routinely attenuates or entirely blocks GNSS signals, visual positioning offers a practical path to accurate machine localisation. The core approach — simultaneous localisation and mapping (SLAM) — processes sequential camera frames to track distinctive visual features (keypoints) across frames, triangulating the camera's trajectory through the environment. When combined with an IMU or wheel odometry, the technique is known as visual-inertial odometry (VIO) and can achieve centimetre-to-decimetre-level accuracy over short distances without any infrastructure. Forest environments present particular challenges because the repetitive texture of bark and foliage can cause feature-matching ambiguities, and lighting varies dramatically between open skies and shaded understorey; researchers have addressed these issues with event cameras, thermal imaging, and learned feature descriptors robust to illumination change. In autonomous harvesting machine research, visual positioning enables a forwarder or harvester to maintain an accurate self-position within a pre-mapped plantation compartment, supporting precise grapple placement and tree-by-tree stem mapping. Australian forestry technology programs, including work at the University of Melbourne's School of Engineering and CSIRO's robotics group, have investigated visual-inertial localisation for field robots operating in eucalyptus and pine plantations. Visual positioning is also integral to UAV-based forest inventory, where downward-facing cameras support hover stabilisation and waypoint accuracy beneath the canopy. As onboard compute becomes cheaper and neural-network-based depth estimation matures, visual positioning is expected to become a standard component of next-generation autonomous forestry machines.
¶ Assessment
Helpful for improving navigation and field efficiency with some safety benefits, but offers limited direct productivity or large-scale risk-reduction impact.
¶ Details
| Field | Value |
|---|---|
| Company | Immersal |
| Country | Finland |
| Type | Company |
| FWPA RD&E | 4.4 |