What SLAM Changed in the Field
The shift from tripod to mobile SLAM is not primarily about speed. It is about which forests we can now capture at all.
We ran tripod LiDAR from our first capture in 2023 through mid-2025. The RIEGL VZ-400i is an instrument of uncommon precision — sub-millimetre range accuracy, 800,000 points per second, a scan sphere that extends to 600 meters in open conditions. For the Elphinstone captures, where we had partner support, clear access routes, and six days on site, it produced 2.4 billion points and a 195-centimetre Douglas fir whose bark texture is resolved to sub-centimetre geometry. That level of detail is irreplaceable for certain uses.
But tripod workflow has an entry cost. Each station requires setup (3–5 minutes), a full scan (6–10 minutes at research-quality settings), and breakdown before movement to the next position. In coastal old-growth — dense understory, uneven terrain, fallen logs, standing water — you cover 1.5 to 3 hectares per day. To adequately capture a 400-acre block, you are looking at weeks of field time.
The introduction of the SLAM+camera unit to our field kit in mid-2025 changed the scale of what we can commit to.
SLAM in Practice
The unit integrates a multi-channel rotating LiDAR with an IMU and a wide-angle camera in a body-worn enclosure. The SLAM algorithm builds and corrects its own positional estimate in real time as the operator walks. There is no station setup, no scan dwell time, no need to carry a tripod into slope terrain. You walk, and the forest is recorded.
Daily coverage depends on terrain and canopy structure. On open forest floor with moderate slope: 12–15 hectares. On steep terrain with dense understory requiring slow careful movement: 6–8 hectares. On areas requiring methodical grid traversal to ensure coverage of dense understory in all aspects: 8–10 hectares.
A two-person team can capture a 400-acre block in 10–14 field days, including GPS control surveys, calibration measurements, and end-of-day processing. The same block with tripod-only workflow would take 6–8 weeks.
What That Unlocks
Speed matters when a block has an active tenure. We have now run captures on five sites where the initial contact from the partner organization was less than three weeks before a scheduled decision date. Under a tripod-only workflow, three weeks is not enough time to get to the site, complete a scan, process the data, and produce deliverables. Under the SLAM workflow, it is barely enough — but it is possible.
Speed also matters for scale. We are a small organization. The reduction in field time per site means we can accept more engagements in a year, which means more forests get captured.
The SLAM workflow has also opened categories of terrain that tripod work made impractical. Very steep ravine slopes — where getting a tripod to a productive scanning position would require rigging — are walkable terrain. Dense cedar grove interiors, where sight lines rarely exceed 15 meters, are traversable. Coastal bluff edges where a tripod setup close enough to capture the cliff face would be unsafe can be covered with the operator standing at the top and angling the body-worn unit.
The Data Integration
We do not treat SLAM data and tripod data as equivalents. In blocks where both have been used — Elphinstone and Avatar Grove — the tripod data anchors the high-precision record of featured specimens, while the SLAM data provides contextual coverage of the surrounding stand.
TreeLearn runs on both datasets, but the accuracy of the SLAM-derived stem measurements is slightly lower (±4.2% DBH RMSE against tape measurements, versus ±2.8% for tripod data on the same stands). We report these differences in deliverables. For stand-level statistics — basal area, composition, old-growth indicator density — the difference in individual-stem accuracy is not material. For individual specimen records that will be cited in a legal filing, we re-scan the specific tree with the tripod unit.
SLAM for coverage, tripod for the specimens that matter most. That is the current standard.