AI/Deep Stack Inquiry

[srvfan]

Ok, I have enjoyed my learning experience with Blue Iris. I think this is an awesome system with so many features I feel as though I will never master it.
One thing I have wondered about is integration with Deep Stack. I get a lot of false alarms on my cameras (insects or birds a lot of times), and I would like to cut down on those. It seems as though AI would be something that could assist in looking for just specific instances/threats.
I know that there is a thread about this subject, but I did not want to hijack the thread with beginner questions. I have skimmed through Wiki (and probably missed it), but is there somewhere I can find information to learn about Deep Stack? Maybe something written for a complete newb as myself?
Are there security concerns with utilizing a program such as DeepStack? I try to be as safety conscious as possible, and I’m worried that maybe facial recognition, personal vehicles, etc may be stored somewhere other than my local computer. I assume that images and videos with Blue Iris are stored only locally, and not accessible by anyone else, but a third party software brings some hesitancy on my part.

 

[wittaj]

Deepstack is downloaded to your computer and ran local.

But I seem to recall you have Andy cameras?

If so, many of us have found that the AI in these cameras are really good. You simply set up BI to pull ONVIF triggers and turn BI motion detection off. I have yet to have a false alarm from my Dahua AI cams. I posted a picture in that thread of my camera AI picking up someone in a blizzard that Deepstack misses. Deepstack when in triggering mode will increase the CPU as well, so depending on your system, it may be an issue.

Deepstack still will get false IDs. People are getting chairs going up their steps and what not LOL. I have a few non-AI cams that I use DS on but would prefer AI cams. Deespstack is certainly getting better though and does fill a niche and with a well defined BI motion detection setup, Deepstack can certainly knock out the few false triggers at that point.

Who uses Dahua AI capable cameras? Reliable AI for triggering events? Pro's/con's?

It has come to my attention that certain Dahua cameras come with built in AI that distinguishes between human & cars. I was unaware of this. Well, always saw the AI word used in the reviews but never really thought about it because I thought it was a Dahua NVR integration only. I am currently...

ipcamtalk.com

 

[srvfan]

Deepstack is downloaded to your computer and ran local.

But I seem to recall you have Andy cameras?

If so, many of us have found that the AI in these cameras are really good. You simply set up BI to pull ONVIF triggers and turn BI motion detection off. I have yet to have a false alarm from my Dahua AI cams. I posted a picture in that thread of my camera AI picking up someone in a blizzard that Deepstack misses. Deepstack when in triggering mode will increase the CPU as well, so depending on your system, it may be an issue.

Deepstack still will get false IDs. People are getting chairs going up their steps and what not LOL. I have a few non-AI cams that I use DS on but would prefer AI cams. Deespstack is certainly getting better though and does fill a niche and with a well defined BI motion detection setup, Deepstack can certainly knock out the few false triggers at that point.

Who uses Dahua AI capable cameras? Reliable AI for triggering events? Pro's/con's?

It has come to my attention that certain Dahua cameras come with built in AI that distinguishes between human & cars. I was unaware of this. Well, always saw the AI word used in the reviews but never really thought about it because I thought it was a Dahua NVR integration only. I am currently...

ipcamtalk.com

Yes sir, I am slowly replacing my existing POS cameras with Lorytas from Andy. I still have several Lorex cameras to go, so I love it when Andy advises of upcoming sales.
I was under the impression that those cameras would only use AI if you had a Dahua NVR. However, I will check out the link you provided and study up on the subject. At least since my camera network is removed from the internet, I have more confidence in images not being stored somewhere other than locally. Thank you.

 

[wittaj]

The cameras AI will work with Blue Iris. What you cannot do prior to DeepStack was do a search by humans or faces with BI. That is the type of stuff you could do with the NVR AI, but you can trigger for those events in BI.

But if all you care about is triggering the camera to send you an alert or push or SMS for human and/or car and do not need all the other bells and whistles that DeepStack offers, this works very well.

How to use the AI in the camera instead of BI motion detection:

Go to Smart Plan and select IVS and hit save.

Then go to IVS and add an IVS rule. Use Intrusion and select the appears and crosses boxes. Draw a box/outline around the area you want it to trigger for, but try not to do the full frame to give the camera time to recognize the object. Then check the box for human and/or vehicle. There is no sensitivity settings in this setup.

Leave min size to 0,0 and do not do any other changes and hit save.

Make sure both motion detection and Smart Motion Detection are not checked in the camera.

Then in BI you need to set it up in two places. In the motion tab, select the camera's digital input or motion alarm.

Uncheck the BI Motion Sensor.

Then go back to the setup screen where you key in the camera IP address and user and password and down in the lower left (or up a little depending on the version you are running), will be the checkbox for pull ONVIF triggers.

I ran a clone camera for awhile using BI motion to compare it to the ONVIF triggers to confirm I wasn't missing anything. Now for those cams I just use ONVIF triggers.

Having the camera do motion detection instead of BI does free up CPU as motion detection is one of the larger CPU drains.

I repeat, do not have Motion Detection and Smart Motion Detection turned on in the camera.

This will eliminate your false triggers.

 

[srvfan]

Man, you are a lifesaver. I was getting ready to add a reply and tell you I would probably be reaching out to you for set up questions. Scary, lol.
For the most part, I think just focusing on humans and vehicles should work for me. I do have one camera where I would like to trigger for the friggin neighbors’ dogs who use my yard for a toilet, but it’s not a huge priority right now.
Thanks for all your help! Geez, maybe one day I will accumulate some of the knowledge you folks have.

 

[wittaj]

Glad to help - feel free to reach out to me if you have any trouble!

 

[JL-F1]

So much agree.

AI with an Andy camera and BI is amazing accurate.

 

[bjohnrini]

@wittaj I repeat, do not have Motion Detection and Smart Motion Detection turned on in the camera.

If I try to disable motion detection while smart motion detection is on, I get this message -

>If you disable motion detection when smart motion detection is enabled, the smart motion detection will be disabled too.

Device TypeSD1A404XB-GNR
System VersionV2.810.0000021.0.R.P9.4a8, Build Date: 2021-01-21

 

[wittaj]

@wittaj I repeat, do not have Motion Detection and Smart Motion Detection turned on in the camera.

If I try to disable motion detection while smart motion detection is on, I get this message -



Device TypeSD1A404XB-GNR
System VersionV2.810.0000021.0.R.P9.4a8, Build Date: 2021-01-21

As it should be. Have both of them disabled and use IVS with human and/or vehicle AI usually results in better operation. Of course field of view plays a role as well. YMMV.

 

[bjohnrini]

As it should be. Have both of them disabled and use IVS with human and/or vehicle AI usually results in better operation. Of course field of view plays a role as well. YMMV.

Oops, I misunderstood what you wrote.
Thanks

 

[srvfan]

Glad to help - feel free to reach out to me if you have any trouble!

Ok, @wittaj, I believe I may have to tag you for your expertise. I have to admit that your suggestion of adjusting the cameras has worked for the most part and has cut down on a lot of false alerts. Thank you! However, I am experiencing an issue with my camera for LPR. Using the same adjustments has cut down on false alarms with bugs and birds. It also successfully picks up people walking into the frame. However, it only picks up vehicles half the time in the day time, and almost zero at night. I’m not sure if it’s because of the angle and zoom I have, but I cannot adjust as it ccaptures plates well. And I do have several small intrusion boxes set up for humans and vehicles as well as tripwires, but still no luck. It’s not a big deal right now, but I would like to see if it can be improved. Thanks

 

[wittaj]

@srvfan - one thing I found with the IVS is that it is very good UNLESS it doesn't have enough time to review the object and determine if it is something to trigger on.

In most use cases, it is not a big deal, but for LPR and how tight we zoom in, then it has troubles for some angles.

And night time is like you said almost zero because the image is so dark so the camera does not see the outline of the car.

So you would have to run simple motion detection for the camera in this scenario, but even then that can be tricky, so you need to run Blue Iris motion detection.

I found this to be the setup that worked best for me. The one zone approach at this location allowed me to catch the front plate of the cars going right to left and the back plate of the cars going left to right. Occasionally I would get a large truck or a very fast vehicle that was missed, but I also run this continuous motion in the event that happens or it missed the motion. And then at night I simply make the object size box a lot smaller like the size of the plate.

 

[srvfan]

@srvfan - one thing I found with the IVS is that it is very good UNLESS it doesn't have enough time to review the object and determine if it is something to trigger on.

In most use cases, it is not a big deal, but for LPR and how tight we zoom in, then it has troubles for some angles.

And night time is like you said almost zero because the image is so dark so the camera does not see the outline of the car.

So you would have to run simple motion detection for the camera in this scenario, but even then that can be tricky, so you need to run Blue Iris motion detection.

I found this to be the setup that worked best for me. The one zone approach at this location allowed me to catch the front plate of the cars going right to left and the back plate of the cars going left to right. Occasionally I would get a large truck or a very fast vehicle that was missed, but I also run this continuous motion in the event that happens or it missed the motion. And then at night I simply make the object size box a lot smaller like the size of the plate.

Roger that. I will take a look at setting up some zones and adjusting the object size once it gets light. Thanks again for the advice/assistance!

 

[biggen]

Are you running an overview cam to compliment your LPR? I have my overview cam trigger my LPR cam. So I use no motion settings at all on my LPR. Its totally dependent on the overview cam detecting motion. Just another option for you if you have trouble fine tuning motion settings on the LPR because of the really small FOV.

 

[srvfan]

Are you running an overview cam to compliment your LPR? I have my overview cam trigger my LPR cam. So I use no motion settings at all on my LPR. Its totally dependent on the overview cam detecting motion. Just another option for you if you have trouble fine tuning motion settings on the LPR because of the really small FOV.

Unfortunately, with my current setup, I do not have an overview cam to mesh with the LPR. However, as I work to replace some of the crap cameras I currently have, I may try to repurpose one to act in that manner. Thank you for pointing out the additional option; I had not considered that before!