Hi there,
I’m working on a method to handle static view obstructions like lightpoles and trees on Telraam V1 on behalf of my bachelor thesis.
Currently I’m testing out how the Telraam report changes when using my method. So I’m running the modified Telraam monitoring algorithm (telraam_monitoring.py) on a test video I created, sending the detection summaries to the Telraam backend by mocking the MAC address. Then I get the corresponding Telraam report via the API on the hourly traffic api.
Unfortunately, the Telraam report’s total number of objects does not match the local summaries I have.
I use a 20 minute test video. The generated summaries contain 486 objects.
I’m expecting the Telraam report to have an uptime of around 0.3333 (20/60 minutes) and the sum of objects (heavy, car, bike, pedestrian) to be around 1.458 (summary objects / uptime = 486 / 0.3333).
But the uptime in the report is only around 0.2969, with 1.312 objects, meaning that only 389 objects made it into the report.
Was there a change in how the backend classifies and calculates the object counts? One month ago it worked well.
Or does anybody have an idea on how to troubleshoot this problem?
Sincerely,
Leander