Robust Single-Image Tree Diameter Estimation with Mobile Phones
Bibliographic record
Abstract
Ground-based forest inventories are reliable methods for forest carbon monitoring, reporting, and verification schemes and the cornerstone of forest ecology research. Recent work using LiDAR-equipped mobile phones to automate parts of the forest inventory process assumes that tree trunks are well-spaced and visually unoccluded, or else require manual intervention or offline processing to identify and measure tree trunks. In this paper, we designed an algorithm that exploits a low-cost smartphone LiDAR sensor to estimate the trunk diameter automatically from a single image in complex and realistic field conditions. We implemented our design and built it into an app on a Huawei P30 Pro smartphone, demonstrating that the algorithm has low enough computational costs to run on this commodity platform in near real-time. We evaluated our app in 3 different forests across 3 seasons and found that in a corpus of 97 sample tree images, our app estimated the trunk diameter with a RMSE of 3.7 cm (R2 = 0.97; 8.0% mean absolute error) compared to manual DBH measurement. It achieved a 100% tree detection rate while reducing the surveyor time by up to a factor of 4.6. Our work contributes to the search for a low-cost, low-expertise alternative to terrestrial laser scanning that is nonetheless robust and efficient enough to compete with manual methods. We highlight the challenges that low-end mobile depth scanners face in occluded conditions and offer a lightweight, fully automatic approach for segmenting depth images and estimating the trunk diameter despite these challenges. Our approach lowers the barriers to in situ forest measurements outside of an urban or plantation context, maintaining a tree detection and accuracy rate comparable to previous mobile phone methods even in complex forest conditions.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".