False Triggers from Lighting Changes - Advice?

[YrbkMgr]

I'm getting an unacceptable number of false triggers due to lighting changes. I'm not sure what settings I should be tweaking so I would appreciate any tips on how to reduce the number of false triggers.

I've included a link to a YouTube video showing the lighting changes - I just did a screen recording while scrubbing the timeline for brevity. In other words, it's not real time, but shows the changes that are triggering events. I've also attached a screenshot of the zone being monitored, if that helps at all.

YouTube Video Link: https://youtu.be/JiKaKf_rmLY

Blue Iris Zone Mask.jpg (212.46 KiB) Viewed 7629 times

[MikeBwca]

You can increase the 'Min. duration (Make Time)' to a value that is longer than the lighting effect, and, still short enough to trigger them walking up to and into your porch. I'd say a value of 1 second should do it.

[YrbkMgr]

You can increase the 'Min. duration (Make Time)' to a value that is longer than the lighting effect, and, still short enough to trigger them walking up to and into your porch. I'd say a value of 1 second should do it.

Thanks a ton for your help. I'll play with that and report back.

[terk]

The only way I could almost completely eliminate false alerts with snow or shadows was spending $50 a year for Sentry AI's smart alerts through BI. Now I only get alerts on people or vehicles on our driveway and people on most of the other cameras, it still records every motion detected it just doesn't alert if it doesn't match what you set per camera and works really well. I only enabled Sentry AI on our outdoor cameras.

[YrbkMgr]

You can increase the 'Min. duration (Make Time)' to a value that is longer than the lighting effect, and, still short enough to trigger them walking up to and into your porch. I'd say a value of 1 second should do it.

So it turns out that after hours and hours of trying to understand what specific terms mean, and I've been using BI since version 3, mind you, I have found success by adjusting trigger parameters. It's not perfect, but I'm making progress...

What I'm doing is using "Test through detector" and then adjusting the triggers minimum object size and "reset detector when object size exceeds", progressively decreasing sensitivity of minimum object size and then varying the reset detector. This is working well, but without definitions, and I mean real definitions of the terms it's a "guess and check" method.

That said, I appreciate your contribution and attempt to help.

[YrbkMgr]

The only way I could almost completely eliminate false alerts with snow or shadows was spending $50 a year for Sentry AI's smart alerts through BI. Now I only get alerts on people or vehicles on our driveway and people on most of the other cameras, it still records every motion detected it just doesn't alert if it doesn't match what you set per camera and works really well. I only enabled Sentry AI on our outdoor cameras.

Thanks for that contribution. What I have found, is when one gets BI and a camera, two things happen: the tendency to get more and more cameras, and the tendency to tweak to "perfection". For myself, I resolved long ago that I don't need continuous recording - I record when there's an alert. So where false alerts are concerned, it's an annoyance that doesn't detract from my main goal of recording when triggered. I've concluded that I don't need to "completely eliminate" false alerts.

I had hoped to arrive at one of two possibilities: prove to myself that BI cannot do what I need without Sentry, or an upgrade to v5 (edge vector algorithm), thus requiring me to evaluate how badly I want to "get there" versus the cost and additional learning curve, OR... learn more about specific settings to allow me to find a satisfactory "sweet spot". I'm all for "purchased solutions", as long as I cannot achieve the same thing by understanding the program parameters better.

All of that notwithstanding, thank you for confirming that the purchased solution works in the event I cannot find my sweet spot.