Seeking Assistance with Eliminating False Positives in Motion Detection

General discussion about Blue Iris
User avatar
Sardon
Posts: 111
Joined: Sun Nov 19, 2023 10:50 am

Re: Seeking Assistance with Eliminating False Positives in Motion Detection

Post by Sardon »

Pogo wrote: Tue Jan 16, 2024 12:01 am FWIW, my 'Make Time' is 0.5.

But that does raise in interesting point. I'm also using a 2MP stream from a camera 150 feet away from the actual trigger zone which is the size of a car at that distance. How relevant that perspective may be compared to a hi-res camera 5 feet away from a comparitively small (almost invisible) zone relative to the likely object motion necessary for trigger activation is certainly an area for consideration. Motion Sensitivity settings (Object Size and Contrast) would certainly be an area worthy of experimentation considering the significant differences between my particular implementation of the method compared to the conditions being discussed here.
I would have thought setting the min. object size to the smallest value of 100 would have the best result considering the zone is so small at 2 pixels across in high res.. This could certainly be a contributing factor
User avatar
Sardon
Posts: 111
Joined: Sun Nov 19, 2023 10:50 am

Re: Seeking Assistance with Eliminating False Positives in Motion Detection

Post by Sardon »

The current make time stands at 1.0. I shall proceed with the necessary adjustments and attempt once more, setting the value to 0.1.

Additionally, while it comes to mind, regarding the clone camera: When I manually initiate an alert on the camera for the purpose of testing my mobile phone's push notifications, which operates as expected due to the settings being cloned, I have observed an issue.

Specifically, if I double-click on the thumbnail of the self-generated alert labelled "external," it displays the following:

error.jpg
error.jpg (6.84 KiB) Viewed 20743 times
So, I am contemplating the implications if we get this functioning correctly and I wish to review the footage swiftly. To elaborate, when motion is detected in my current setup, it generates alerts on the Balcony camera, and all I do is double-click on the alert to play the footage from that point. Therefore, in this scenario with a cloned camera generating alerts via JPEGs, I am wondering if I will be able to click on the alert quickly and have it play the footage from that point?

In any case, the first step is to make the adjustment and see if this resolves the issue. Fingers crossed, it could well be that the make time setting is causing an issue; a duration of 1 second might not have been sufficient to trigger the alert.

Right, I'm off to adjust the make time value, test it, and then I'll report back to you all.
User avatar
Pogo
Posts: 389
Joined: Tue Jul 18, 2023 7:21 pm
Location: Reportedly in the Area

Re: Seeking Assistance with Eliminating False Positives in Motion Detection

Post by Pogo »

Sardon wrote: Tue Jan 16, 2024 12:05 pm I am wondering if I will be able to click on the alert quickly and have it play the footage from that point?
The alert from the clone will only be a jpeg of the trigger activity as being discussed here unless it is also (or exclusively) set to record motion as well. But that's what the main camera is for.

I simply refer to the main camera live view when receiving an alert requiring additional observation. Othwerwise the jpegs are usually adequate for my purposes.
User avatar
Sardon
Posts: 111
Joined: Sun Nov 19, 2023 10:50 am

Re: Seeking Assistance with Eliminating False Positives in Motion Detection

Post by Sardon »

It's fantastic to report that I had a delightful surprise when testing the camera. Just as I was about to dispose of the rubbish and having adjusted the make time to 0.1, I had a timely visitor at my door. A quick look at my watch even showed a thumbnail of the postman!!

clone-alerts.jpg
clone-alerts.jpg (18.2 KiB) Viewed 20733 times
I'm really excited about the next few days as I'll be monitoring the system to see how well it performs under various lighting conditions etc. Now that the clone setup is up and running, I'm thinking of hiding the camera from Blue Iris interface, as it is currently distorting the other camera feeds.

Also when I double-click on the postman's thumbnail, I get a 'No File Found' message, just like I mentioned before. It's not a big deal, but I do wonder if a static JPEG is meant to be displayed ?

As for managing alerts, I'll now need to explore the a new approach I guess. In the past, as mentioned , I'd simply double-click an alert to view the footage. I'm considering whether I should now use the time bar for manual navigation, although I have ever rarely used it and have even removed it from my screen in increase real-estate for my cameras. I'm also pondering how to review alerts after hiding the clone? Is there a method in which I can do so ?

The progress made, particularly with Pogo's trip wire suggestion, is great. I must also extend a special thanks to TimG for pinpointing the make time issue – a one-second delay was clearly excessive.

The next step for me is to learn a new method for managing and reviewing footage triggered by alerts. Given that my cameras record non-stop, I might focus on simply being aware of when an alert is generated and manually finding the relevant footage.

Overall, I'm thrilled with these developments – it's all moving in a very positive direction!
User avatar
Sardon
Posts: 111
Joined: Sun Nov 19, 2023 10:50 am

Re: Seeking Assistance with Eliminating False Positives in Motion Detection

Post by Sardon »

Pogo wrote: Tue Jan 16, 2024 12:21 pm
Sardon wrote: Tue Jan 16, 2024 12:05 pm I am wondering if I will be able to click on the alert quickly and have it play the footage from that point?
The alert from the clone will only be a jpeg of the trigger activity as being discussed here unless it is also (or exclusively) set to record motion as well. But that's what the main camera is for.

I simply refer to the main camera live view when receiving an alert requiring additional observation. Othwerwise the jpegs are usually adequate for my purposes.

Hey Pogo I was busy typing out the the good news and missed your message!

Your explanation is crystal clear. Regarding the settings, do you reckon I should maintain the minimum object size at 100, given the high resolution and contrast size at 20? I'll consider adjusting them only if the lighting conditions trigger a false alarm, though that might not be necessary. It seems that with these settings, the system is primed to detect any movement crossing the trip wire zone, which is quite remarkable!

I'm really excited, lets see how well this holds up to changing lighting conditions!

Again, massive thank you to you guys, couldn't have done this without your support which is greatly appreciated
User avatar
Pogo
Posts: 389
Joined: Tue Jul 18, 2023 7:21 pm
Location: Reportedly in the Area

Re: Seeking Assistance with Eliminating False Positives in Motion Detection

Post by Pogo »

Congratulations!

You may now benefit from reviewing the thread for other overlooked details which could further assist your refinement of the setup's performance.

BTW... ;)
Honic Clone Motion Sensor.jpg
Honic Clone Motion Sensor.jpg (89.16 KiB) Viewed 20728 times
User avatar
Sardon
Posts: 111
Joined: Sun Nov 19, 2023 10:50 am

Re: Seeking Assistance with Eliminating False Positives in Motion Detection

Post by Sardon »

Pogo wrote: Tue Jan 16, 2024 1:02 pm Congratulations!

You may now benefit from reviewing the thread for other overlooked details which could further assist your refinement of the setup's performance.

BTW... ;)
Honic Clone Motion Sensor.jpg

I must admit, I completely missed the importance of the Make-Up time setting, Pogo. That was quite an oversight on my part, apologies for that.

In a rather surprising twist, the placement of my trip wire is proving to be quite eventful. You see, the flat below mine, when the heating is switched on, causes condensation exactly where the trip wire is situated. It's almost hard to believe – I've already received alerts because of this!
damn.jpg
damn.jpg (44.18 KiB) Viewed 20695 times
I'm thinking it might be wise to slightly shift the trip wire to the left. What do you think?

On another note, now that I'm not getting alerts on the main camera, I'm wondering how to swiftly access and review the footage. At the moment, double-clicking the camera just maximises it to full screen. Is there an efficient way to view recorded footage without needing the timeline constantly displayed at the bottom of the screen? Any tips would be greatly appreciated!
User avatar
Sardon
Posts: 111
Joined: Sun Nov 19, 2023 10:50 am

Re: Seeking Assistance with Eliminating False Positives in Motion Detection

Post by Sardon »

No worries at all, I've just realised I can simply click on 'all clips' at the top left, select them, and then scrub through the footage to the time indicated by the alert notification. Brilliant!

I've also decided to increase the thresholds to match yours :) Let's see how that works out!
User avatar
Sardon
Posts: 111
Joined: Sun Nov 19, 2023 10:50 am

Re: Seeking Assistance with Eliminating False Positives in Motion Detection

Post by Sardon »

I've just discovered something else. When I hide the clone and choose alerts, then turn off the 'solo' option, the alerts for the clone disappear. This means I can't check them unless I go into 'edit layout', choose 'show hidden cameras', and then they re-emerge.

I wonder if there's any clever workaround to display the alerts? As it stands, this method seems quite time-consuming for navigating alerts, which is something I frequently did in my previous setup.

It seems I'll have to adapt to this, although it's already proving somewhat inconvenient.

Another possibility is to hide the main camera and focus on the clone, but then I face the same issue when wanting to review actual recorded video footage. Or, I have to deal with, what was for me, the tricky process of adjusting the layout. I'm not sure why Blue Iris introduced this 'Edit Layout' feature; the windows used to align seamlessly before, but now positioning them is rather cumbersome. I could try to fit all cameras on screen, but it feels a bit odd having two identical cameras occupying screen space.
Attachments
solo.jpg
solo.jpg (829 Bytes) Viewed 20676 times
User avatar
Sardon
Posts: 111
Joined: Sun Nov 19, 2023 10:50 am

Re: Seeking Assistance with Eliminating False Positives in Motion Detection

Post by Sardon »

In fact I've just had an idea..

Now that we know this works. I could take a backup from my main camera and apply the exact same configuration to the main camera and delete the clone.

Surely this would be the best way to address all issues.
Post Reply