IPC-T5442T-ZE IPC-T5442TM-AS IPC-T5842T-ZE SMD 3.0 Smart IR Latest New Firmware From EmpireTech

Yes it’s currently set to 0, 0 but whenever a fly or with heavy rain it triggers the IVS as I have cars parked on my driveway and I have person and car detection enabled.

Interesting. I haven't experienced that with my AI cams (the non-AI cams yeah bug and rain trigger all night).

Post a video of it happening and from the SD card if you have one installed so we can see the IVS rules and the outline when it triggers.

I mean mine didn't trigger during a snow storm except for when someone was out walking

1642374989040.png
 
Yes it’s currently set to 0, 0 but whenever a fly or with heavy rain it triggers the IVS as I have cars parked on my driveway and I have person and car detection enabled.
Can you confirm what firmware version/release on you on, specifically? Many of us that are concerned about IVS, seem to focus on using the Dec 2020 release firmware for the 5442 cameras which has had the most consistent results (IMHO).
 
  • Like
Reactions: user8963
Can you confirm what firmware version/release on you on, specifically? Many of us that are concerned about IVS, seem to focus on using the Dec 2020 release firmware for the 5442 cameras which has had the most consistent results (IMHO).

i have a 98% hitrate with dec2020 day and night with ir light only from camera and some street lights.

smd 3.0 fails too much for me.
human filter is ok for ivs with dec2020, but not as good as deepstack. sometimes it got triggered by something else. but no fail alarms from rain, reflections or else.

so +1 for dec20 :smoking:
 
Last edited:
  • Like
Reactions: EMPIRETECANDY
We invite @looney2ns to test the fw, Here is the testing result from him. This firmware not support Smart IR, but we will add on it in Feb.
So far, the IVS is working very well. I believe this firmware has also solved the problem of the video pausing every second I have seen the past few firmwares.

If anyone want to try, just download here.
 
Can you confirm what firmware version/release on you on, specifically? Many of us that are concerned about IVS, seem to focus on using the Dec 2020 release firmware for the 5442 cameras which has had the most consistent results (IMHO).

I’m running the Dec 2020 firmware, have tried most of not all the versions since that Andy has posted but have always gone back to the Dec version.
 
  • Love
Reactions: user8963
Wow surprised you're having the issues on the Dec 2020 firmware. Can you upload/send/share an example video for us to look at that shows these false positive triggers?
 
Also double check that regular motion detection hasn't been turned on somehow. I've had that happen a couple of times to different cams. I know I didn't change it but sure enough when I went to look it was switched back on again.
 
Sorry for the delay but I waiting for another false trigger, got one just a few minutes ago, you can see something quickly buzzing across the bottom right which is what I assumed triggered IVS.

 
Can you share that same video but including the IVS intrusion box or tripwire (in blue) on the video as well? That will show us for sure if something was triggered by that bug in the beginning of the video, but without IVS drawn in, there's no certain way to know for sure. You should be able to get that of your camera (if it has an SD card) or from the NVR itself.
 
Sorry for the delay but I waiting for another false trigger, got one just a few minutes ago, you can see something quickly buzzing across the bottom right which is what I assumed triggered IVS.



did you calibrate the camera?

In my opinion your mounting is somehow wrong (too high, too sharp angle, strange twisted image) and it is to be expected that IVS/SMD will fail. So calibrate can maybe help a bit
To my knowledge these "pointing down"-mountings are always problematic.

Are you using tripwire or intrusion ?
Intrusion could be problematic, because the area is so small caused by the "pointing-down"-mount so the camera might have trouble to detect if something enters the box. Intrusion needs free space and works best
 
Last edited:
I don’t have an SD card in the cam, yes I know I’ve been meaning to add one but have never got round to doing it.

I did roughly calibrate it ages ago, by this I mean I had no idea what I was doing but I have been reading about this again and might give it another go.


My house is on a corner which is why the pavement/sidewalk curves away.

That is mounted in the furthest point possible on the corner of the house, any further and the cam would be in the road height is around 9-10 ft

I use both intrusion and tripwire, both set to within the boundary of the driveway.
 
  • Haha
Reactions: user8963
The cam went crazy last night from between 3am and 5am with an alert alert very few minutes. Checked the clip and there’s really nothing that I can see. When I’ve had this before a restart normally fixes the issue for a while.

Also noticed that the BI alert images are missing but that another issue?

IMG_8936.jpg
 
I had a number of IVS triggers last night from a small object in a spider web (see video). This is on one of the cameras (5442 series) running the new FW 11-05-21 with Intrusion detection box. I originally had the sensitivity set to 6, because I do not get any animal intrusions in this screen enclosure. I set the sensitivity to 5 and increased the minimum size to at least 2 times the size of this circular blob, and still got false IVS triggers (See insert) . That was a little disturbing, since I have filtered out animals in the past with min size objects. I personally feel that you should never keep min size at 0,0. Yes, this camera is properly calibrated with Global Setup. All light is from camera IR and external IR light. no visible light.

On a positive note I have had very good results with this FW version during rain storms, tree canopy moving shadows and fog. FW does have algorithm problems discriminating animals and "blobs moving around. Overall, I am happy with the IVS performance, but feel the algorithms are still in their infancy and need some improvement. Hopefully, some of this information gets back to the Dahua FW guys, so they can make future improvements. Someone else stated that "you are expecting too much". That may well be true, but hopefully things will get better. If people don't provide more feedback on their experiences, changes may be slow in coming.


1642777403465.png

 
  • Like
Reactions: sebastiantombs
Yeah, I would make the intrusion box smaller. Also looks like you have WDR on at a high level? That can cause problems too.
 
Boy, eveyone has an assumption. The intrusion box triggered is on the right. I usually try to not have any of the tripwire/boxes overlapping. All backlight is Off for Night setting. I am not a novice for setting camera. I set the IR lighting on the high side, so I can increase shutter speed at night.

1642780157700.png
 
  • Like
Reactions: sebastiantombs
Well since you aren't a novice then I guess we will stop giving advice LOL :lmao:

You said you had a problem and we were offering up suggestions based on the information provided by you...

With that said I am hard headed and will offer up one more suggestion - have you tried min object size of 0,0 and let AI do it's thing? Or is that insulting and you know that and already tried that :lmao:

All of mine are set to 0,0 and I do not get these types of false triggers. You should only use a min object size if you have like a dog that is being mistaken for a human for example.
 
  • Love
Reactions: Flintstone61
It is not exactly disabled (at least I do not think), but when a size is put in, it will then be "more willing" to trigger for an object larger than that. Keeping it at 0,0 is supposed to force it to rely strictly on AI.

Do you have an SD card in the camera that you can playback with IVS lines drawn to confirm that it was an IVS trigger and not that somehow MD got turned on (if you use BI some of have seen MD get turned on in the camera with a reboot of the camera).
 
  • Like
Reactions: Flintstone61
No, I do not use SD cards in my cameras. I even went back and checked, there is not any MD enabled in the camera or BI. I can see from the original video that the triggers happened when they crossed the intrusion box. Not to start any animosity, but I have searched the internet about IVS in the past and have never found any documentation about setting min size emphasizes AI, one way or another. Is this speculation? Could you provide a pic of the scene you have?
 
Last edited: