[Partially Solved] Lame IVS on IPC-T5442TM-AS

[tigerwillow1]

(Edited to add findings after 7 weeks of testing and experimenting. Initial post is immediately below)
The IVS on my T5442TM-AS won't pick up anything but large objects. It picks up people and deer, while dogs and cats walk right through the tripwires and intrusion boxes without detection. By comparison, all my other cameras reliably pick up small birds and chipmunks. I figure it has to be either defective firmware or operator error. The firmware version is 2.800.0000000.10.R . On the operator error angle, I've tried all sorts of different IVS settings and exposure settings, as well as factory reset. On the other cameras IVS has been easy and works well. Can't get this one to work. Are any of you having better luck with it? Is there anything obvious that I'm missing?

(Findings after 7 weeks)
1. The IVS problems I have with the T5442TM-AS are apparently not widespread, based on reports from others.
2. Mine uses a 6mm lens which Andy does not stock. No prospect of a replacement until coronavirus supply disruption has ended.
3. I'm able to use the camera with firmware version V2.800.0000000.8.R . With other versions, it's a brick to me.

Definition of "IVS doesn't work" for this post: Tripwire and intrusion will not detect objects the size of cats, small dogs, foxes, etc. People and vehicles are detected in all the firmware versions I've tried.

Firmware version results:
V2.800.0000000.1.R: IVS works great. Camera sometimes does not send the IVS End Event message causing NVR timeline to show a constant IVS event, which is just as useless as never showing an IVS event.

V2.800.0000000.8.R : IVS works acceptably with its sensitivity set to +10. Detection not as good as with the .1.R version, but combined with a little bit of motion detection it does the job I need it to.

V2.800.0000000.10.R and V2.800.0000000.11.R : IVS doesn't work, as defined above. Useless for my situation.

Long term plan: Because the low light sensitivity is so darn good with the T5442TM-AS I'll probably buy another one when the supply line is running again. With the two of them side-by-side I can verify if the camera I have is obviously defective. If it is, I'll ask Andy to replace it.

 

[bigredfish]

Sure you have the minimum size box small enough?

 

[tigerwillow1]

I started out with the minimum size at 0-0, and tried a couple other small values on the chance there is a boundary condition or initialization error in the firmware. No luck with any of that. Also turned on and off the people and vehicle filters in case of initialization bug.

 

[bigredfish]

Hmmm mine seems to be the opposite. Get plenty of cats and coons... mines a bullet though

 

[tigerwillow1]

And are you running the 2.800.0000000.10.R firmware? If you are, maybe I should post screen shots of my settings.

 

[Buddym]

I have the same trouble, some of my 5442 works pretty fine and some didn't.
To Set those cams to factory default an implementing IVS from scratch doesn't solve it.

I have installed 10.R.

 

[looney2ns]

Let @EMPIRETECANDY know, so he can continue to provide feedback to Dahua. These cams have had a tough history with buggy firmware out of the gate.

 

[aristobrat]

I swear I saw a post about someone going into other tab on the IVS page (Global Setup) and using some wood strips laid out in the cameras view to calibrate things and then IVS working a lot better. I'll see if I can find that post.

 

[aristobrat]

Here we go. Different camera model, but I'd assume it's still applicable.

Just to let you know my experience. I have a 5241 running System Version V2.800.0000000.8.R, Build Date: 2019-09-02. It's mounted just under the roof overlooking a driveway. We have lots of cats and every night they used to trigger the IVS tripwires - at least 5 times. Since running a calibration (3 vertical, 1 horizontal) it's been perfect. Not one misfire in a week. The gDMSS notification to my LG mobile, just say either "human" or "vehical". Works brilliantly for me!.

and

DahuaWiki

dahuawiki.com

This is the guide I followed. Very simple. I just got 3 * 1m lengths of wood and stood them up on the drive to calibrate the verticals, and laid 2 out horizontally to calibrate that axis. Makes sense really as without any reference to what the camera is looking at, how would the algorithms know the relevant sizes of objects detected. The multiple messages on gDMSS are due to having multiple tripwires. As I say, after a week of multiple cat triggers literally every couple of hours, I know only get humans or vehicle

 

[looney2ns]

I swear I saw a post about someone going into other tab on the IVS page (Global Setup) and using some wood strips laid out in the cameras view to calibrate things and then IVS working a lot better. I'll see if I can find that post.

View attachment 53711

Good catch, my snot filled head knew that too, but couldn't be retrieved.

 

[bigredfish]

Im running V2.800.0000000.10.R, Build Date: 2019-11-18 on my bullets

 

[Markc_UK]

I swear I saw a post about someone going into other tab on the IVS page (Global Setup) and using some wood strips laid out in the cameras view to calibrate things and then IVS working a lot better. I'll see if I can find that post.

View attachment 53711

That was me...lol. Still working great since calibrated.

 

[tigerwillow1]

Let @EMPIRETECANDY know, so he can continue to provide feedback to Dahua.

Agree with this. I'm just trying to first determine if the issue is related to one specific firmware or if there's a workaround. I'll try the calibration procedure in the next day or two. I don't want to waste Andy's time with a report that doesn't have enough information or might just be operator error. With all the forum members who have installed 5442s, I'm surprised there isn't a larger quantity of solid feedback.

 

[Markc_UK]

Here's a daily example (apologies for the quality, I've zoomed right in). One of the neighbours just loves to round up his (many!) cats and you can see IVS distinguishing between them. The camera is armed with IVS rules that play warning sounds during throughout the night, which is hasn't falsely triggered (as yet).

 

[IAmATeaf]

Good catch, my snot filled head knew that too, but couldn't be retrieved.

Try blowing it

 

[FlipNJ]

Here's a daily example (apologies for the quality, I've zoomed right in). One of the neighbours just loves to round up his (many!) cats and you can see IVS distinguishing between them. The camera is armed with IVS rules that play warning sounds during throughout the night, which is hasn't falsely triggered (as yet).

Recently got help for my camera and was surprised how much using highlight control and basic image adjustment helped. the second cat passes along the dark curb or railroad tie so there is not enough of a contrast difference for it to catch it. The reflected light off the driveway surface is probably messing up the trip. Try HLC and see if it helps. The IVS on Dahuas is extremely tricky.

 

[tigerwillow1]

The IVS on Dahuas is extremely tricky.

I've not found this to be the case with my other cameras, but I'm paying attention to every suggestion. Even my chinese market cameras have great IVS detection. Many times, it's such a low contrast moving image that I have to play the clip a few times to spot it. I did the IVS calibration and changed another setting on the 5442, but now the animals aren't cooperating. We're having some ugly weather and the cats, dogs, coyotes, foxes, skunks, and porcupines I usually see are apparently hunkered down. Even our own cats refuse to go out.

 

[Markc_UK]

Recently got help for my camera and was surprised how much using highlight control and basic image adjustment helped. the second cat passes along the dark curb or railroad tie so there is not enough of a contrast difference for it to catch it. The reflected light off the driveway surface is probably messing up the trip. Try HLC and see if it helps. The IVS on Dahuas is extremely tricky.

Hi. I think there may be misunderstanding. Prior to calibrating the camera IVS was triggered by everything that crossed its path. As there are many cats locally, the system was triggering literally every hour. Now it only triggered if a human or vehicle passes. Works great for me.

 

[tigerwillow1]

Hi. I think there may be misunderstanding.

Probably good you brought that up. I was aware of this and almost posted that I was looking for the opposite and want the cats to trip the IVS. I did the calibration just because I'm trying anything and everything. It makes sense to me that the calibration could be useful for deciding if objects are humand or vehicles. This morning a buck walked right through the intrusion box without detection. I'm starting to suspect that human detection might be permanently stuck on. Just a theory.

 

[looney2ns]

Probably good you brought that up. I was aware of this and almost posted that I was looking for the opposite and want the cats to trip the IVS. I did the calibration just because I'm trying anything and everything. It makes sense to me that the calibration could be useful for deciding if objects are humand or vehicles. This morning a buck walked right through the intrusion box without detection. I'm starting to suspect that human detection might be permanently stuck on. Just a theory.

Might try this, do a factory reset to defaults at least 3 times. Others have noted this sometime fix's weird behavior in these cams.