Detecting without Training: An Open-Vocabulary Object Detection Method for Identifying Hazardous Objects on Construction Sites
Bibliographic record
Abstract
The construction site is one of the most hazardous workplaces, accounting for more than 20% of worker fatalities in the United States. The leading causes of these injuries are falls, slips, and trips, which contribute to over a third of the fatalities in the construction industry. Therefore, identifying hazardous objects on the construction site that might lead to these injuries is crucial for ensuring the safety of construction workers. Many previous studies have explored resolving these safety issues by monitoring either the construction workers or their working environment. However, most of these previous studies encounter practical challenges, primarily due to cameras being fixed, which resulted in difficulty in detecting objects that are obstructed or hidden. In this study, we propose a novel framework that uses open-vocabulary object detection (OVOD) based on large vision language models, which could efficiently detect potential hazardous objects on the construction sites without training. We collected more than 5,000 egocentric images of real construction sites in Tuscaloosa, Alabama, as the data set for validating our framework compared to other popular classification network architectures. The results show that the proposed framework can successfully identify hazardous objects on construction sites, which include outdoor and indoor working environments, with over 79.0% weighted F1-score compared to the best performance of 75.7% achieved by other traditional classification methods that were trained on our data set. The proposed framework has the potential to help safety managers improve the safety conditions on construction sites.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
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 source (direct Gemma or distilled Codex), 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".