IVS option missing

[aristobrat]

IMO, that's a great pic!

 

[Fastb]

Here's an example of an object too large to trigger a tripwire.

yes, that object box is large.
However, when the truck came into view, farther down the driveway, and "acquired" as an object to track, the box was smaller. Giving the camera time to acquire the object, with a smaller box.
IMHO, the trip lines could be moved, so the truck would break the lines before the box grew in size. More on that in a bit.
Another thought: The lines are placed so that only when the truck was close would the lines be crossed.
Another observation: The lines are short, so only a portion of the big box crosses the line. Some of the box wouldn't cross the complete line, only a portion of the line. In the pic, a part of the box is already outside "Line 2" tripwire. That portion would be excluded from the "50% line cross" calculation, making it harder to trigger an event. As the truck moves up the driveway, and the box moves to the right, even more of the box is excluded from the 50% calculation.
Similarly, for "Line 4" it doesn't seem the box would ever meet the 50% rule. Too much of the box would pass above "line4".
Conclusions:
- I suggest using several vertical trip lines, running from bottom to top of the frame. Maybe just a bit inside the frame edges.
The small box, (when the truck is further away), would cross a well placed line. When getting closer, the larger box would cross another line or two.
Maybe the vertical lines could be slanted, say from 11:00 to 5:00. This might reduce the time it takes to meet the 50% threshold.
- IVS could be made to work it seems to me, by experimenting with different lines that were placed carefully, with knowledge of box movement direction and box sizes (from small to large).
- This could help when the truck leaves the property. Right now, with the current lines, by the time a leaving truck is acquired and "boxed" (this takes some time), the box already is outside the 50% threshold. And won't meet a 50% rule as it travels down the driveway....

Good luck with experiments!

Fastb

 

[tigerwillow1]

On the Penske truck picture, the biggest challenge I have is for everything coming up the driveway to trigger an event, while at the same time having nothing just passing by on the road cause a trigger. This has been quite a challenge, and that's why all of the trip lines are squished at the bottom of the frame. Even with this, I get false triggers at night from the headlight glare of passing cars, or when the road is wet, headlight reflections off of the pavement. I never tried long vertical lines, so that's a good suggestion. I'm skeptical, but sometimes I get surprised, too. The camera is where it's at to catch license plates coming up the driveway. That's also a challenge. I can adjust the exposure so it works good in daylight, or at night, but a common setting for both has eluded me so far. I just swapped out a chinese market 4mp camera for the 2mp starlight hoping that will help. Results TBD.

 

[Kinchyle]

See if this helps on gmail notifications to phone. You don't need vpn. I would recommend blocking your nvr from any outbound ports except the one needed.
Not sure if anyone saw the FBI advisory on Dahua and Hikvision. Should be ok if you are safe with firmware upgrades and segment your network

How to get NVR or camera alerts to your phone notification status bar

 

[ss205c]

I can adjust the exposure so it works good in daylight, or at night, but a common setting for both has eluded me so far. I just swapped out a chinese market 4mp camera for the 2mp starlight hoping that will help. Results TBD.

Is there no day / night profile on that camera so you can set different settings for each? I guess I presumed all the dahua's had that.

 

[tigerwillow1]

The cameras have always had day and night profiles but it's only with relatively new firmware that the profiles will switch automatically when the camera switches mode due to ambient light. With the current version 2.640.0000002.0.R firmware I'm able to have different day and night exposure settings and this has made for an overall improvement in reading license plates. But there's now a new issue: When the automatic day/night profile switching is enabled, the setting for the photocell sensitivity is disabled. So now, when the ambient light is pretty low, the camera is in day mode (set for automatic exposure) where it runs a really long shutter time making for blurred motion. I'm going to try a high tech solution for this: Tape over the photocell.

For detecting vehicles on the driveway, this is the one and only case where I'm having better luck with motion detection over IVS.

 

[bababouy]

Any body having issues with IVS actually triggering SmartPSS? We have several cameras that are triggering Smart PSS and some that just show up in the log as dissappeared. They are set up in the recorder properly and when an object crosses the line, they turn red and flash like they are supposed to. The trigger is read by Smart PSS, but the pop up doesn't come up and no audio alert.

 

[cuz]

yes, that object box is large.
However, when the truck came into view, farther down the driveway, and "acquired" as an object to track, the box was smaller. Giving the camera time to acquire the object, with a smaller box.
IMHO, the trip lines could be moved, so the truck would break the lines before the box grew in size. More on that in a bit.
Another thought: The lines are placed so that only when the truck was close would the lines be crossed.
Another observation: The lines are short, so only a portion of the big box crosses the line. Some of the box wouldn't cross the complete line, only a portion of the line. In the pic, a part of the box is already outside "Line 2" tripwire. That portion would be excluded from the "50% line cross" calculation, making it harder to trigger an event. As the truck moves up the driveway, and the box moves to the right, even more of the box is excluded from the 50% calculation.
Similarly, for "Line 4" it doesn't seem the box would ever meet the 50% rule. Too much of the box would pass above "line4".
Conclusions:
- I suggest using several vertical trip lines, running from bottom to top of the frame. Maybe just a bit inside the frame edges.
The small box, (when the truck is further away), would cross a well placed line. When getting closer, the larger box would cross another line or two.
Maybe the vertical lines could be slanted, say from 11:00 to 5:00. This might reduce the time it takes to meet the 50% threshold.
- IVS could be made to work it seems to me, by experimenting with different lines that were placed carefully, with knowledge of box movement direction and box sizes (from small to large).
- This could help when the truck leaves the property. Right now, with the current lines, by the time a leaving truck is acquired and "boxed" (this takes some time), the box already is outside the 50% threshold. And won't meet a 50% rule as it travels down the driveway....

Good luck with experiments!

Fastb

Wait a minute, isn't IVS supposed to be easier than just using motion detection?
Or is it so much more complicated to allow you to customize to a much greater depth than just using motion detection?
The info above is making my head hurt...

 

[tigerwillow1]

Wait a minute, isn't IVS supposed to be easier than just using motion detection?

For me, not a simple yes or no answer. I don't think IVS is any easier to set up. Once it's set up, that's when things get a lot easier in the form of a dramatic elimination of false positives. With all of my cameras outdoor, motion detection gives so many false positives that it's close to useless. I imagine there are some indoor situations where motion detection works great.

For the situation I posted earlier ( the Penske truck coming up the driveway), IVS has failed me and it's the one camera out of 14 where I'm using motion detection. I think one of the IVS functions could be made to work here, but the way it's implemented, it doesn't. Yet I'm still a strong defender and supporter of IVS. This one camera is in a very uncommon situation, looking down a driveway that parallels the road with about 10' of separation. I have it oriented to read license plates, and it's just darn hard to trigger on driveway traffic without also triggering on road traffic. With both IVS and motion detection, I get a lot of false triggers when it's set up to get the triggers I really want. I'm getting the "least worst" ratio with motion detection. When it snows, the motion detection becomes useless.

With all that qualification, I'll stand up and say IVS is overall easier and more reliable, with an occasional exception in an unusual setup.

 

[cuz]

For me, not a simple yes or no answer. I don't think IVS is any easier to set up. Once it's set up, that's when things get a lot easier in the form of a dramatic elimination of false positives. With all of my cameras outdoor, motion detection gives so many false positives that it's close to useless. I imagine there are some indoor situations where motion detection works great.

For the situation I posted earlier ( the Penske truck coming up the driveway), IVS has failed me and it's the one camera out of 14 where I'm using motion detection. I think one of the IVS functions could be made to work here, but the way it's implemented, it doesn't. Yet I'm still a strong defender and supporter of IVS. This one camera is in a very uncommon situation, looking down a driveway that parallels the road with about 10' of separation. I have it oriented to read license plates, and it's just darn hard to trigger on driveway traffic without also triggering on road traffic. With both IVS and motion detection, I get a lot of false triggers when it's set up to get the triggers I really want. I'm getting the "least worst" ratio with motion detection. When it snows, the motion detection becomes useless.

With all that qualification, I'll stand up and say IVS is overall easier and more reliable, with an occasional exception in an unusual setup.

Thanks for that info. Just to clarify, are you doing any of the IVS setup in the cameras or is it all being done at the NVR?

 

[aristobrat]

Any body having issues with IVS actually triggering SmartPSS? We have several cameras that are triggering Smart PSS and some that just show up in the log as dissappeared. They are set up in the recorder properly and when an object crosses the line, they turn red and flash like they are supposed to. The trigger is read by Smart PSS, but the pop up doesn't come up and no audio alert.

There have been a couple of folks reporting issues like this. Sometimes the forgot to enable IVS in the Smart Detect area of the camera, sometimes they didn't have IVS recording schedules setup... sometimes everything as correct and they just needed to reboot to initially get it going. :/

 

[tigerwillow1]

Thanks for that info. Just to clarify, are you doing any of the IVS setup in the cameras or is it all being done at the NVR?

Another not a simple answer for me, thanks to the jungle of firmware versions and plugins. In most cases, I set up IVS on the camera. But a couple of my cameras don't play well with the plugins I have and I do the setup on the NVR. Some NVR functions work on Pale Moon and not Firefox, with the opposite for some other NVR functions. I'd use the NVR for IVS setup all the time except it doesn't work with a few of the cameras. Ironically, the chinese market cameras give me the least hassle in this regard. As I updated NVR firmware versions over time I noticed that IVS setup on the NVR worked better with each revision. I can't upgrade any more because of the newer firmware actively rejecting the chinese market cameras.

The dahua NVR is doing the job I need it for and has been very reliable, but frankly, knowing what I know now I'd go with BI if I were starting over.