1. I would like to see how AI seeks to address false alarms caused by nighttime insects flying close to the cameras and crawling on the domes (e.g. moths). Perhaps a screened enclosure with a few IPVM-trained moths (ha ha).
2. For those of us that live in bear country, can Hanwha's (or others') AI settings be adjusted so specifically to allow alarming on animals but filter out other false alarms?
Josh, thanks! We're planning a separate test of that plugin and others. There is a Genetec plugin and integration to Hanwha Wave/Nx Witness. At the time of this test there were some issues and limitations to integrations and not everything was working properly, which Hanwha expected to be addressed in the next couple of weeks.
Well, I guess that is progress. But "...BriefCam edge analytic initiative is underway with availability targeted for H1 2021", with no specifics yet, to compare to existing offerings, not to mention that they are way behind. That's coming from a guy that loves Axis and uses almost nothing else.
What were the results of object classification at night or low light levels? IR on IR off? Distance? Forced to stay color or B/W?
Classification performed the same at night using the integrated IR of the camera and detected at the same PPF as it would in day. I just walked around in a parking lot in ~4lx (measured on me) with the camera forced to day mode and it dropped the detection to ~22 PPF (~90' with the fov at its widest) so you do lose detection distance in a lower light setting in color mode (distance shown below).
Noise overload AI & Metadata push to server?
Are you asking if noise in the scene caused false alarms? If so, I did not see any of that during testing.