Better understanding needed of 'wait until triggered at least x times in y s'

[Dave Lonsdale]

I'm a long standing user of BI but have always suffered with too many false motion alerts. Extending the minimum duration, masking, cancel shadows, using IR away from the camera lens etc helps but is only partially effective. And now with more than 20 cameras I'm having another go to reduce them.

Disappointingly, setting all cameras to 'wait until triggered at least 2 times in 10s' appears to have made no difference for sunlight changes (that cause a vast number of false triggers). When I review the alert video footage, a single red alert rectangle will track a cloud shadow across the screen for >10s, although its shape does change within the 10 seconds. Can anyone please expand on BI's help text?

 

[fenderman]

I'm a long standing user of BI but have always suffered with too many false motion alerts. Extending the minimum duration, masking, cancel shadows, using IR away from the camera lens etc helps but is only partially effective. And now with more than 20 cameras I'm having another go to reduce them.

Disappointingly, setting all cameras to 'wait until triggered at least 2 times in 10s' appears to have made no difference for sunlight changes (that cause a vast number of false triggers). When I review the alert video footage, a single red alert rectangle will track a cloud shadow across the screen for >10s, although its shape does change within the 10 seconds. Can anyone please expand on BI's help text?

for sunlight use reset when exceeds...you need to use multiple settings to achieve accuracy including zone crossing.

 

[Dave Lonsdale]

Dear Fenderman, thanks for your quick reply. I have to be cautious restricting the maximum image size. Most cameras have a distance view where sunlight changes cover a relatively small area but need to capture near objects. I'll have another go with zone crossing - to date it hasn't been effective (unlike Dahua's IVS built into newer cams which I've found to be brilliant).

But I ask again, should BI trigger in the presence of a single but changing red rectangle when ...2 times... is set?

 

[fenderman]

Dear Fenderman, thanks for your quick reply. I have to be cautious restricting the maximum image size. Most cameras have a distance view where sunlight changes cover a relatively small area but need to capture near objects. I'll have another go with zone crossing - to date it hasn't been effective (unlike Dahua's IVS built into newer cams which I've found to be brilliant).

But I ask again, should BI trigger in the presence of a single but changing red rectangle when ...2 times... is set?

if you have two triggers within 10 seconds it will..any two triggers will set it off...if you properly setup line crossing as per the help file and threads discussing them it will be better than dahuas line crossing which is very limited in comparison....also understand that the goal is not only to eliminate false triggers but to make sure you get all the actual triggers..

 

[Dave Lonsdale]

OK fenderman, I'm really pleased that BI's line crossing is better than Dahua's. I must have screwed up when I trialed it previously. May take a while, but I'll compare and report back.

 

[aristobrat]

May take a while, but I'll compare and report back.

Interested in hearing how it goes for you.

As pleased as I am with BI, I still haven’t been able to tune its Motion Detection to be as good with ignoring light changes with things like rolling shadows (from fast moving clouds on sunny days) as Dahua’s IVS is on my Starlight turrets.

 

[Dave Lonsdale]

Hello again fenderman. I have been trialling multiple zones through last week but need your expertise. I have used my copy of BI at home with a single camera covering the drive, set up twice in BI, one having the basic motion sensor with a single zone and the other using zones A through G.

Please look at the attachments showing the settings which are identical except one of them used only zone A (zones box unchecked) and the other with zone A masked and zones B to G using bands moving progressively away from the camera down the drive.

You will see from the snapshot of my wife walking the dog down the drive that zones B, C and D do not trigger, whereas the other having one zone (greenery masked) does trigger. To show this more clearly, I have included another one with the massive object of a car driving slowly by. Zones B, C and D sense but do not trigger. The single zone has a full red rectangle until the car has passed through.

I assume that if I use a single large zone to cover the area of B, C and D then I will get triggers but in doing so would defeat the object of trying to reduce false triggers with multiple zones.

So fenderman, are you able to help please? I am anxious to do something. For example, on this sunny and breezy day, on my main BI installation, 17 cameras created 840 triggers - all false! Needless to say, I don't send alert emails etc, triggers are reviewed using remote web access on iPad/iPhone.

 

[anijet]

From the BI help file:

".....you don't want to leave masked "gaps" between zones when looking for objects crossing zones. Overlapping the zones is not ideal either because the object may originate in the overlapped region and will never be detected as crossing."

 

[fenderman]

dave see anijet's post above..go back an reread the help file..

 

[Dave Lonsdale]

Mmm yes, thanks again for the instant replies. In my defence, I had read the help file but mistakenly thought that "One way to configure this would be to use zone A as a mask, excluding areas that you do not want to monitor at all. Then, setup additional zones B and C that are part of A, and look for object movement between B-C. This way, B and C do not have to be touching or overlapping whatsoever."...... meant that the zones no longer needed to touch each other. I suppose this means that zone edge boundaries adjacent to masked zone A will not trigger - ie only one of four sides as in the helf file example.

 

[Dave Lonsdale]

Bingo! I made the zone bands B to G across my drive touch each other and continued with the comparison test through this afternoon. The scene was very difficult - bright sunshine and windy giving dancing shadows (snapshot attached). Over a three and a half hour period the result was as follows:-

BI with a single zone (greenery masked) 74 false triggers
BI with 6 touching zone bands plus zone A masked 4 false triggers
Dahua IVS (greenery masked) 44 false triggers
Both real triggers successfully recorded on all all three setups

Remarkable! I still can't really believe it and will continue the comparison test overnight. Could do with some insects, rain, fog, bats or whatever (integral IR in this bullet camera is switched on). The challenge will then be to create zones for all the other cameras on the main system that have far more awkward views to cover with touching zones and are almost all PTZs, each having many presets.

Once again, very many thanks to the forum contributors.

PS I still think the help file is misleading.

 

[Dave Lonsdale]

OK, in conclusion after experimenting endlessly with various settings, I unfortunately have to report that I have been unable to repeat the remarkable reduction in what would appear to have been a freak weather-related false trigger result of 3rd July enjoyed when using multiple object crossing zones within BI. The reduction subsequently has been typically one third.

Moreover, even having set the Object travels to only 50 pixels in 1080p/2MP video, the zone setup misses pigeons, blackbirds, rabbits and badgers strutting around on my drive that are captured by both BI without crossing zones (all other settings identical) and Dahua's IVS.

Nevertheless, BI's software is so user friendly, especially when using the Web app for both alerts and real time monitoring, that it still wins the day - even when there are many false triggers.