I don't think so, fluffy shows on the captures, but not the lady. Also I never knew VLC would play the .dav files
Dahua WizColor 5x59-PRO and SmartLight 5x59-IL new series
- Thread starter steve1225
- Start date
I don't think so, fluffy shows on the captures, but not the lady. Also I never knew VLC would play the .dav files
Correct me if this is wrong:
ROI doesn’t apply to IVS at all – only to motion detection or exposure. For IVS, the rule itself acts as the ROI, and the AI still analyzes the full frame before checking whether a bounding box actually crosses the line or zone.
Global Smart Illumination can indeed trigger the white LEDs independently of IVS. If the IVS log didn’t show an event but the light came on, that’s likely what happened.
So disabling the global AI illumination and using IVS-based Linkage Actions instead isolates this behavior perfectly.
Increasing IVS sensitivity (8–9) usually increases false triggers on wide FOV outdoor scenes, since bounding boxes get larger.
So I’d try:
Disable global Smart Illumination
Use IVS event linkage for light control
Keep sensitivity moderate (3–4)
That combination should stop the random triggers from people far outside the Tripwire while keeping accurate IVS reactions.
ROI doesn’t apply to IVS at all – only to motion detection or exposure. For IVS, the rule itself acts as the ROI, and the AI still analyzes the full frame before checking whether a bounding box actually crosses the line or zone.
Global Smart Illumination can indeed trigger the white LEDs independently of IVS. If the IVS log didn’t show an event but the light came on, that’s likely what happened.
So disabling the global AI illumination and using IVS-based Linkage Actions instead isolates this behavior perfectly.
Increasing IVS sensitivity (8–9) usually increases false triggers on wide FOV outdoor scenes, since bounding boxes get larger.
So I’d try:
Disable global Smart Illumination
Use IVS event linkage for light control
Keep sensitivity moderate (3–4)
That combination should stop the random triggers from people far outside the Tripwire while keeping accurate IVS reactions.
bigredfish
Known around here
There was no random trigger. Smart Illumination did what it was supposed to do and no IVS event was recorded.
Bounding boxes don’t trigger IVS rules. They ID potential targets and categorize them.
The actual object then has to cross or interact with the IVS line/rule and typically takes 1/2 second or so for AI to verify and trigger the rule
Lets look at one
Notice IVS is not triggered by the bounding box, its triggered when the human target actually crosses the line
4 frames go by after the target hits the line, time for AI to validate and trip
No Activation by bounding box
Activation by Human target
View attachment Recording 2025-10-09 125806.mp4
A few things from my point of view.
My pictures were both set up with exposure 0-1 IR mode. So yeah, this cam is as good as the S3 cams at the moment, and hopefully with updated firmware, even better.
The 3 Pro can also trigger the LED lights from IVS setup. Never bothered to look at event linkage. haha.. but the 3 pro doesn't trip for animals, if there are people looking for that setting.
The bullet 5 pro has longer IR because it's a 8-32mm lens, vs the turret at 2.7-8mm.
I personally prefer bullet as well, but I don't really need the bullet for the back or side yards.
My pictures were both set up with exposure 0-1 IR mode. So yeah, this cam is as good as the S3 cams at the moment, and hopefully with updated firmware, even better.
The 3 Pro can also trigger the LED lights from IVS setup. Never bothered to look at event linkage. haha.. but the 3 pro doesn't trip for animals, if there are people looking for that setting.
The bullet 5 pro has longer IR because it's a 8-32mm lens, vs the turret at 2.7-8mm.
I personally prefer bullet as well, but I don't really need the bullet for the back or side yards.
bigredfish
Known around here
This one may be even more telling
Car bounding box does not trip IVS tripwire in first two photos
It takes almost 42 frames or 1.3 sec for the AI on this older camera to verify the target as a vehicle and trip the event (from the bounding box first touching the tripwire on exit)
View attachment SloMoRecording 2025-10-09 132022.mp4
Car bounding box does not trip IVS tripwire in first two photos
It takes almost 42 frames or 1.3 sec for the AI on this older camera to verify the target as a vehicle and trip the event (from the bounding box first touching the tripwire on exit)
View attachment SloMoRecording 2025-10-09 132022.mp4
Last edited:
Ok, still trying to figure out what turned on the led warm light. I do not use Smart Illumination during my night profile. Also it was the only time it came on when she walked by.
That difference in perspective has a big impact on how accurately the AI interprets the Tripwire interaction.
bigredfish
Known around here
I'll try and find the thread as its months ago, but I was able to prove that with IVS disabled in the 3449H-AS-PV-PRO, the global scene Smart Illumination still triggered the white LED's to come on. Its activated much like SMD anywhere in the FOV within reason
There is no Event Linkage on the 3449Pro (well the fixed version anyway) to connect IVS to activating the white lights. Only the Red/Blue warning lights
Sounds like the 54PRO has the Event Linkage in addition
bigredfish
Known around here
Sorry wrong image
Do you have other IVS rules that night besides the one on the steps? Did you by chance have MD?SMD active?
There is no way that sidewalk tripwire picked up that lady in the street imho
Last edited:
But that is exposure, not Illuminator, where the Smart Illumination setting resides.
You would think a IVS event would show in log...nothing. Was hoping to find a clue there.
bigredfish
Known around here
Thats why I say its wasnt IVS that tripped the light. It would also theoretically show up in the recording as an IVS event
bigredfish
Known around here
MD/SMD on?
Maybe the car headlights reflection if you play it back slow?
But only two things can active the lights I think.
Maybe the car headlights reflection if you play it back slow?
But only two things can active the lights I think.
nope, no MD or SMD. But now I am concerned why the led lights did not come on fluffy. I might just default the cam back to factory defaults and start fresh. I tried Andy's 09/28 fw and then went back. I did the default, but now wondering if something is hosed ? I'll report back with better results hopefully.
View attachment Recording 2025-10-09 125306.mp4
@Tinman,
just curious — when the warm LEDs came on that night, even though Smart Illumination was off in your profile.
Would be interesting to see if you notice a one-second light pulse or a quick flicker right before the person leaves the frame.
just curious — when the warm LEDs came on that night, even though Smart Illumination was off in your profile.
Would be interesting to see if you notice a one-second light pulse or a quick flicker right before the person leaves the frame.
I guess, but maybe Andy can tell more?
The camera’s AI is tuned to treat “Human” classifications with higher confidence than “Animal.”
So when it detects a person, it may briefly activate the warm LEDs as an assist light, even without Smart Illumination or a full IVS trigger.
But when it sees an animal, especially at night and with lower contrast, the classification confidence is lower — so the AI won’t trigger any lighting response.
That’s why the lady caused the LEDs to flash, while “Fluffy” didn’t. It’s not a fault. Just how the WizMind model weights object types differently.
bigredfish
Known around here
There is a quick flash/reflection from the car headlights before you actually see the car
I posted this in the "T54PRO-ZE versus 5442-S3 low-light" thread, but figured it's worth cross-posting here since it includes some daytime footage.
While the T54PRO-ZE is much better than the 5442-S3 (and the shutter speed is 3X faster!) in (this!) low-light situation, there's not a whole lotta difference between them during the day (with good light!) ... and I actually like the saturated colors of the 5442-S3 a bit more - both are set to 60% saturation - I think the White Balance (both are set to Auto) is being handled differently.
While the T54PRO-ZE is much better than the 5442-S3 (and the shutter speed is 3X faster!) in (this!) low-light situation, there's not a whole lotta difference between them during the day (with good light!) ... and I actually like the saturated colors of the 5442-S3 a bit more - both are set to 60% saturation - I think the White Balance (both are set to Auto) is being handled differently.
Here is yesterday morning. Note no car headlights this time. It does appear the led light pulses some as she is leaving the frame. Also note, I have the warning light to stay on for 20 seconds, well it stayed on for 65 seconds.
bigredfish
Known around here
I think thats just resetting exposure after the light went out. I see it on the 5442's even worse