A Vision-Based System Design and Implementation for Accident Detection and Analysis via Traffic Surveillance Video
Bibliographic record
Abstract
In this work, we aim to investigate the problem of detecting and analyzing traffic accidents automatically and effectively through surveillance videos and implement the whole framework on an AI demo board. First, the technique of motion interaction field (MIF) that has the potential to detect crashes in a video is adopted to locate the crashed vehicles based on the interactions between multiple moving objects. Second, the YOLO v3 model is employed to identify the crashed vehicles within the appropriate location. In order to recover the vehicle trajectories before the collision, a hierarchical clustering approach is used, and the corresponding trajectories are obtained. Third, to facilitate the judgment of traffic police, the trajectory is projected to a vertical view by using a perspective transformation. The vehicle velocity is estimated accordingly with the unbiased finite impulse response (UFIR) approach that does not require statistical knowledge of the external noise. Then, the estimated velocity and the obtained collision angle from the vertical view can be utilized to analyze the traffic accident. Finally, to show the effectiveness and implementation performance of the proposed approach, an experiment is carried out based on a Huawei AI demo board named HiKey970 that is used for coding all the mentioned algorithms. Several accident surveillance videos act as the input of the demo board. The accidents are detected successfully, and the corresponding vehicle trajectories are recovered.
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 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 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 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".