Do Blue Iris have something similar to Dahua IVS? AI require online connection?

TechBill

Known around here
Joined
Nov 1, 2014
Messages
1,770
Reaction score
1,175
My cameras and NVR are Dahua brand. I been using it for about 5-6 years now. It still work great but the mobile app.. it's so-so eh ... not great but not bad either.

I am thinking of looking into Blue Iris to use for mobile notification however all my camera being Dahua 2MP Starlight camera, I use IVS triplines to alert me of any motions in area and it been very reliable for me with very little false triggers. I believe somewhere I did read up that Blue Iris does not support any IVS trigger from Dahua cameras.

My question is does Blue Iris have it own version of IVS which work with Dahua camera or any brand camera?

I haven't read much about AI person detect or Deepstack but does it require online connection and using cloud server to analysis the videos? Does it require higher end computer to be able to use those features in Blue Iris?

Thank you
 

sebastiantombs

Known around here
Joined
Dec 28, 2019
Messages
11,511
Reaction score
27,690
Location
New Jersey
Blue Iris does support ONVIF from Dahua cameras or any other ONVIF compliant camera, for the most part. I use it along with motion detection in Blue Iris and segregate which is which by using a clone camera for one or the other. The easy way to test this is to download the trial version of Blue Iris and give it a go on that. If you're satisfied you can convert that to a licensed version so you don't lose any of the work put it while trying it out.

DeepStack and the newer one, SenseAI, are both run locally on your machine. DeepStack currently supports CPU and GPU processing versions. SenseAI supports CPU processing only but they are working on a GPU version as well. Custom models are also supported by both. A model is the collection of objects you want to detect. The "base" model that comes with each has a lot of objects we normally don't look for, like a PC, table, giraffe, elephant, zebra, well you get the idea. By limiting the number of objects to what you're really looking for vastly improves detection time which, in turn reduces processing load on either the CPU or GPU.
 

TechBill

Known around here
Joined
Nov 1, 2014
Messages
1,770
Reaction score
1,175
Blue Iris does support ONVIF from Dahua cameras or any other ONVIF compliant camera, for the most part. I use it along with motion detection in Blue Iris and segregate which is which by using a clone camera for one or the other. The easy way to test this is to download the trial version of Blue Iris and give it a go on that. If you're satisfied you can convert that to a licensed version so you don't lose any of the work put it while trying it out.
I don't use motion in Dahua .. just the IVS

I could be wrong but I thought IVS is not registered in ONVIF complaint, just the standard motion detection. I use mostly tripline IVS detection for my alerts which I think is very reliable for me. I will try the trial version on an old computer. I use Macbook Pro however I do have Windows install on it. I will try that first to test it out before spending money on a computer for Blue Iris.
 

TechBill

Known around here
Joined
Nov 1, 2014
Messages
1,770
Reaction score
1,175
Blue Iris does support ONVIF from Dahua cameras or any other ONVIF compliant camera, for the most part. I use it along with motion detection in Blue Iris and segregate which is which by using a clone camera for one or the other. The easy way to test this is to download the trial version of Blue Iris and give it a go on that. If you're satisfied you can convert that to a licensed version so you don't lose any of the work put it while trying it out.

DeepStack and the newer one, SenseAI, are both run locally on your machine. DeepStack currently supports CPU and GPU processing versions. SenseAI supports CPU processing only but they are working on a GPU version as well. Custom models are also supported by both. A model is the collection of objects you want to detect. The "base" model that comes with each has a lot of objects we normally don't look for, like a PC, table, giraffe, elephant, zebra, well you get the idea. By limiting the number of objects to what you're really looking for vastly improves detection time which, in turn reduces processing load on either the CPU or GPU.
Thanks for the clear explanation so it does sound like I need a good high end computer for local processing then.

Is it on the same machine as Blue Iris or it need to be on a seperate machine with a seperate processor?
 

sebastiantombs

Known around here
Joined
Dec 28, 2019
Messages
11,511
Reaction score
27,690
Location
New Jersey
The horsepower of the BI machine is dictated by how many cameras are running AI. More cameras mean higher horsepower. That can be mitigated by using a GPU for processing AI. Add in using sub streams in BI and horsepower becomes even less important. Rule of thumb is a generation 8 or better i5or i7 processor with 16GB of RAM and a CUDA capable GPU. A system like that can easily handle 30, or more, cameras quite easily. The problem is that GPUs have gotten pretty expensive compared to what they used to be. Typically a used card will run at least $150, and up, depending on which card you actually settle on. Anything from a GTX1030 on up will work well.
 
Last edited:
Top