Three of the pipelines behind Simvera are open for evaluation. Request access, receive an access token by email, then upload a frame from your own cameras and read the result, with a written reason attached to every finding.
Your email, your role and what you want to test. Takes a minute. Nothing is uploaded yet.
A person at Simvera reads it and emails you an access token. You will not see it on this page.
Choose a pipeline below, enter your email and access token, and upload an image.
An annotated image plus the numbers and reasons behind it. Runs take upt to a minute per image.
Each pipeline runs on ordinary camera frames. Sample outputs below come from our own test footage; your images are processed the same way and never used for anything else.
Finds restricted floor areas marked with tape, cones, chains, barriers or painted stripes, fills the enclosed floor as a zone and marks the boundary with a warning ring. Then measures how far each person stands from that boundary, in metres. Uses the camera pose when it can estimate one, or a top-down reading when the view is overhead, and can check those people for the PPE the zone requires.
Catalogues the scene, rates each hazard with a danger probability and a written justification, outlines the ones above threshold, and checks every person nearby for the protective gear that hazard requires.
Segments the drivable road, then anything on or beside it that is not a vehicle: debris, rocks, potholes, animals, fallen objects. Works from dashcams and roadside cameras, day or night.
Pick a pipeline, add an image, run. Each run spends one attempt. Results stay available for six hours.
Enter your access token to run this pipeline, or request one if you do not have one yet.