CodeProject.AI Version 2.5

It's looking like CPAI can't identify vehicles from a high-mounted camera. This image returned "Nothing found" from all of the models. The Dahua IVS correctly identified it as a vehicle. The same truck was identified correctly from 2 other cameras mounted a few feet off the ground.
UpperGully.20240131_140830.2464204.5-0.jpg
 
  • Like
Reactions: David L
For now I have the whole enchilada of what's loaded by default: license-plate, ipcam-animal, ipcam-combined, ipcam-general, ipcam-dark. But does that have anything to do with the AI status saying something was identified at 60%, and the BI log saying nothing found?
Not sure but what I did is Clone each Camera I wanted a certain model to run on. It just made it easier for me to do that. Not sure if having several models slows things down...I know there is a reason there are different models...I am sure Mike or someone will chime in with the right answer.
 
  • Like
Reactions: tigerwillow1
It's looking like CPAI can't identify vehicles from a high-mounted camera. This image returned "Nothing found" from all of the models. The Dahua IVS correctly identified it as a vehicle. The same truck was identified correctly from 2 other cameras mounted a few feet off the ground.
View attachment 184819
In AI tab in BI Settings do you have Use custom models checked and Default object detection unchecked?
 
It's looking like CPAI can't identify vehicles from a high-mounted camera. This image returned "Nothing found" from all of the models. The Dahua IVS correctly identified it as a vehicle. The same truck was identified correctly from 2 other cameras mounted a few feet off the ground.
View attachment 184819
It can certainly detect that truck...see below both using defaults accuracy gets better using custom ipcam-general at 86%
 

Attachments

  • truck2.PNG
    truck2.PNG
    239.6 KB · Views: 87
  • truck.PNG
    truck.PNG
    239.2 KB · Views: 86
  • truck 3.PNG
    truck 3.PNG
    243.5 KB · Views: 75
  • Like
Reactions: 105437 and David L
In AI tab in BI Settings do you have Use custom models checked and Default object detection unchecked?
I did not have Default Object Detection unchecked and will try it that way. What is it for? There's nothing about it in the help file or manual.
 
I did not have Default Object Detection unchecked and will try it that way. What is it for? There's nothing about it in the help file or manual.
Don't know just was told when using custom-models to uncheck it...if checked I don't think custom-models work...Mike, Fenderman can correct me here...
 
just was told when using custom-models to uncheck it...if checked i don't think custom-models work...Mike, Fenderman can correct me here...
My custom models worked with this checked, except for one camera. It would be nice to find out what it's for.
 
Don't know just was told when using custom-models to uncheck it...if checked I don't think custom-models work...Mike, Fenderman can correct me here...
The custom models do work. I used to use the default however with ipcam-general that only detects humans and vehicles i get much faster times 130ms on average. Albeit with a few false alerts per week - something I rarely saw with the default.
 
  • Like
Reactions: 105437 and David L
I have a few more things to try. On my 3 cameras that work ok with AI, I use only ONVIF triggers from the camera, and Static Object Analysis is checked. On the one that always says Nothing Found I also have the BI Motion Sensor active and for some reason had Static Object Analysis unchecked. From reading I don't think the Static Object Analysis checkmark should matter, but I'll try to figure out why one camera works different from the others.

fenderman, thanks for running the image offline through CPAI. Is that something that's easy to do by myself?
 
  • Like
Reactions: David L
I have a few more things to try. On my 3 cameras that work ok with AI, I use only ONVIF triggers from the camera, and Static Object Analysis is checked. On the one that always says Nothing Found I also have the BI Motion Sensor active and for some reason had Static Object Analysis unchecked. From reading I don't think the Static Object Analysis checkmark should matter, but I'll try to figure out why one camera works different from the others.

fenderman, thanks for running the image offline through CPAI. Is that something that's easy to do by myself?
Very easy...open the cpai interface by going to in the browser. Then click on codeproject ai explorer (highlighted in blue). You will then see the screen in my screenshots. Use detect objects for default and custom detect to use a specific model - which you can choose from the dropdown.
 
  • Like
Reactions: 105437 and David L
Are any suggestions for fixing the failed detection of a cat? It's not the confidence setting in BI. All of the models in CPAI returned "Nothing found". The cat is near the right edge at the bottom of the house. The camera's IVS drew a box around it and tracked it no problem.
View attachment 184763

Here's an interesting analysis. For my purposes I don't care if it's the correct animal or not, just that it's an animal.
View attachment 184764
You'd probably do best with @MikeLud1 's ipcam-dark GitHub - MikeLud/CodeProject.AI-Custom-IPcam-Models for the dark image, but the cat is pretty small so your mileage will vary.

And that's totally a dog. The rare slender legged large ear'd kelpie, if I'm not mistaken.
 
Very easy...open the cpai interface by going to in the browser.
I got the same truck 69% that you did. I'm getting to suspect more and more that there's some interaction with the ONVIF alert and BI's motion detection but will experiment more before sticking my foot deeper into my mouth.
 
  • Like
Reactions: David L
And that's totally a dog. The rare slender legged large ear'd kelpie, if I'm not mistaken.
It's definitely a feral black cat. I catch it on other cameras and it comes up to our door sometimes.
 
I got the same truck 69% that you did. I'm getting to suspect more and more that there's some interaction with the ONVIF alert and BI's motion detection but will experiment more before sticking my foot deeper into my mouth.
Yes, the same image will yield the same result. You need to look at what images blue iris is sending to CPAI and how its returning results. Hold ctrl key while double clicking on the alert and it will display the blue iris status ai analysis page. Note you must have "save ai analysis details" checked in the AI settings for the camera.
 
  • Like
Reactions: tigerwillow1
Hold ctrl key while double clicking on the alert.......
On the left is the jpeg from the BI Alerts directory, on the right is what it sent to CPAI. This explains non-detection of the truck.

InAlertsDirectory.jpg SentToAi.PNG

The blacked out area is what I have excluded for BI's motion detection, to not trigger on cars on the road. The alert says it was triggered by ONVIF alerts, so I wouldn't expect BI to apply the motion detection mask. I'd also expect the jpeg in the Alerts folder to be exactly what it sends to CPAI. Maybe there's some logic here I don't understand? I'm thinking I can maybe get around this by cloning the camera. I haven't learned cloning yet, need to learn it tomorrow. Looks like the BI watermark isn't sent to CPAI either.
 
  • Like
Reactions: David L
On the left is the jpeg from the BI Alerts directory, on the right is what it sent to CPAI. This explains non-detection of the truck.

View attachment 184850 View attachment 184851

The blacked out area is what I have excluded for BI's motion detection, to not trigger on cars on the road. The alert says it was triggered by ONVIF alerts, so I wouldn't expect BI to apply the motion detection mask. I'd also expect the jpeg in the Alerts folder to be exactly what it sends to CPAI. Maybe there's some logic here I don't understand? I'm thinking I can maybe get around this by cloning the camera. I haven't learned cloning yet, need to learn it tomorrow. Looks like the BI watermark isn't sent to CPAI either.
This is an older post when I started using Delivery model on cloned camera.


You will have a Master and the clones from each camera are slaves...
 
Hi

I have BI setup with approximately 15cams running (Hikvisions) some of these are clone cams for AI alerts. I am detecting 'person'only using ipcam-combined.

Is this the best model to use with YOLOv5.Net 1.9.2? or is there a faster way to process?

I typically set to detect 5 images at 500ms

I notice then when triggered and detecting CPU spikes to 100% for a few seconds and then drops down to around 15%

Is the spiking normal or is there something I can do to assist this. I have a Lenovo AIO pc with 16gb Ram

CPU - i5-10400T @ 2.00GHz
6cores 12 processors
Intel UHD 630 GPU


Status logs in CPAI report around 250ms -1500ms at time or higher

Also get Random shutdowns

Use Main Stream is unticked

Any advice/help is appreciated

thanks
 
@ch3st3rk Given the details you shared, since you unticked main stream, BI5 should send substream image to CP.AI. Is substream configrued for that camera?

Also, please share screenshots for:
#1: AI tab for that camera
#2: The AI details for a triggered event.
 
What's a good spec to aim for -windows PC - to run this plus blue iris plus video storage for five Dahua cams please?
Ideally I'd like a mini pc to keep the power usage down. Apologies if this is an annoying question for the thread. I have the cameras which were storing to internal chips but we'd like an upgrade and some AI so the pc will be bought for this purpose.