CodeProject.AI Version 2.0

[Pentagano]

I have zero mqtt knowledge. Total noob, I can do BI, pushover, even manages a Kasa light alert with an exe by jdfromdb, but that's it. So I would have to start any home auto from scratch, which is why I suggested that I knew a tutorial for the sprinklers would be unlikely, but worth asking anyway. I might someday try to learn all this, but too big a task for now.

It is comprised of many components - mqtt push to my broker (in BI Alerts mqtt). Nodered flow receives this from the broker listening (red arrow)
Blue arrow is a filter - each dot an object cat/dog/person for example
Pink - just a switch if I want the flow to continue
Green - linked to my sonoff device which turns it on or off.
You could achieve something similar all in homeassistant though.

 

[Pentagano]

I use Node-RED for my notifications and it works well.
View attachment 168080

The function node is:


I like it because you're not restrained in what you can do compared to configuring alerts in a UI. It's just JavaScript code so you can easily modify the code to do whatever you want.

How does the tv one work? You get an image that pops up on the screen?
I have this flow also but only to my phone. Added this and pushover after my telegram was down for 2 days the other week!

 

[JL-F1]

What I'm seeing with the critters model is that it likes to identify flying insects at night as four-legged animals. In this case a dog.
View attachment 168130

I am having the same problem with 'critters'. bug is 70% dog/cat, and this is with bugs close to the camera so zoning would be hard

 

[Pentagano]

I am having the same problem with 'critters'. bug is 70% dog/cat, and this is with bugs close to the camera so zoning would be hard

Are you running in B/W with IR mode also?

 

[JL-F1]

Are you running in B/W with IR mode also?

Color, no IR. Annke NCD800 which has it's own lights

 

[carteriii]

@carteriii can you share how you got YOLOv5.NET to use your built-in Radeon graphics? is there some settings that I am missing? My PC has a Radeon RX Vega M GL Graphics (and I can see it under CodeProject.AI's System Info) but when click Enable GPU my YOLOv5.NET still shows "CPU".

@Palmpilot I didn't do anything special, so I don't really have any specific instructions for you. HOWEVER, I did find elsewhere in this forum and perhaps even this thread that someone else said the "CPU" is still displayed for some period of time until several executions have successfully executed with the GPU. If it's not working, I believe there should be some indication in CodeProjectAI logs. Be aware that changing from CPU to GPU also takes some time, effectively stopping & restarting the service. You should see that in the logs as well. At this point, I don't remember if I explicitly restarted the service or not, but I've certainly restarted numerous times since then so the setting/selection remains.

Obviously there are other possible problems. I haven't looked up your card, but make sure it supports DirectML and that you have a current driver. It's also possible that the manner in which CodeProjectAI is detecting or using the GPU simply isn't (yet) supported for your card. I would hope something in the logs would give you an indication if there's a different problem. Only because you didn't explicitly mention the logs, be sure you're looking there first for any problems before trying anything else (like a driver update).

I don't know if it's helpful, but here's a screenshot of what I see, with lovely performance times AND 1.4M executions since my last restart.

 

[Pentagano]

Color, no IR. Annke NCD800 which has it's own lights

OK guess you just have bigger bugs than me unfortunately

 

[JNDATHP]

What does DirectML mean?

 

[MikeLud1]

What does DirectML mean?

 

[JNDATHP]

When I switched to YOLOv5.net, this is what I see. I have an Nvidia RTX 2060. Am I using the Nvidia card for YOLOv5.net?

 

[MikeLud1]

When I switched to YOLOv5.net, this is what I see. I have an Nvidia RTX 2060. Am I using the Nvidia card for YOLOv5.net?

View attachment 168681View attachment 168682

Everything looks normal

 

[Pentagano]

Cat detection. Difference between deepstack on medium and CPAI with critters model (large) v6.2.
Deepstack higher percentage but wrong animal. Timing was almost the same.

 

[David L]

When I switched to YOLOv5.net, this is what I see. I have an Nvidia RTX 2060. Am I using the Nvidia card for YOLOv5.net?

View attachment 168681View attachment 168682

You are getting good Detection times too...

 

[dubber]

I'm currently using IVS rules. I have a few IVS rules where one section of the field of view is only for person, and the other is set for person and cars.

Is there a way to set up BI Motion and CPAI to achieve similar results?

Example: Zone 1 will be only setup for person detection and Zone 2 will be setup for both person and cars?

If the above is not possible, is there a way to setup just one zone to detect only person?

 

[David L]

I'm currently using IVS rules. I have a few IVS rules where one section of the field of view is only for person, and the other is set for person and cars.

Is there a way to set up BI Motion and CPAI to achieve similar results?

Example: Zone 1 will be only setup for person detection and Zone 2 will be setup for both person and cars?

If the above is not possible, is there a way to setup just one zone to detect only person?

You may consider Cloning:

CodeProject.AI Version 2.0

It would be interesting to see some benchmarks with different systems, running two instances Because if with one instance it takes 100ms and with two takes 200ms, you have gained nothing...

ipcamtalk.com

 

[dubber]

You may consider Cloning:

CodeProject.AI Version 2.0

It would be interesting to see some benchmarks with different systems, running two instances Because if with one instance it takes 100ms and with two takes 200ms, you have gained nothing...

ipcamtalk.com

I have two spotter cams

 

[Pentagano]

You may consider Cloning:

CodeProject.AI Version 2.0

It would be interesting to see some benchmarks with different systems, running two instances Because if with one instance it takes 100ms and with two takes 200ms, you have gained nothing...

ipcamtalk.com

I use the AITool with Deepstack and CPAI with BI.
VorlonCD/bi-aidetection: Alarm system for Blue Iris based on Artificial Intelligence. (github.com)
Specify the below directory in the AITool.
With deepstack gpu or cpu installed, the AItool has options to start the deepstack and models/size etc etc.
CPAI controlled by BI and deepstack by AITool.
Works great for me and serves as a backup/refinement

I

 

[MikeLud1]

I'm currently using IVS rules. I have a few IVS rules where one section of the field of view is only for person, and the other is set for person and cars.

Is there a way to set up BI Motion and CPAI to achieve similar results?

Example: Zone 1 will be only setup for person detection and Zone 2 will be setup for both person and cars?

If the above is not possible, is there a way to setup just one zone to detect only person?

See below, from BI help page 84

 

[dubber]

See below, from BI help page 84

View attachment 168714

Thank you for the reply.

My current setup is below. To confirm only have "Person" but every time a car pass by, my spotter cam would get triggered.

Am I suppose to put person:80 in the "To confirm" area?

 

[Pentagano]

Thank you for the reply.

My current setup is below. To confirm only have "Person" but every time a car pass by, my spotter cam would get triggered.

Am I suppose to put person:80 in the "To confirm" area?

View attachment 168715

Yep this is what I have.