[tool] [tutorial] Free AI Person Detection for Blue Iris

joshwah said:
Anyone know what these errors mean when compiling?


Code: 
1>------ Build started: Project: UI, Configuration: Debug Any CPU ------
1>CSC : error CS1617: Invalid option '8.0' for /langversion. Use '/langversion:?' to list supported values.
========== Build: 0 succeeded, 1 failed, 0 up-to-date, 0 skipped ==========
Click to expand...
Anyone know?
 
I am running AI Tool 2.0.319.7666 from 12/27/2020. Everything had been working well and I got the detections working the way that I want. However, this morning I went to check and AI Tool had shut down overnight. I have restarted it and rebooted the machine multiple times. AI Tool runs and the queue fills up with 100 images, but none of them are ever processed. It just stays at "Working".

1611348008923.png

This seems to me to be an issue with Deepstack? But it is running:
1611348110067.png

These are the errors in the log. (Network path is correct now)
1611348172509.png

How can I get this working again? I have re-installed by overwriting the files in the AI Tools directory with the originals.

Edit to add that I have manually tested Deepstack with a POST command and it is running and working properly.
 
Last edited:
MrPie said:
My program was working very well but a few days ago I noted it says there were many errors, so I checked the log and I can see that sometimes it tells me that:
"the process cannot access thephotofile.jpg because it is being used by another process"
And then
"ERROR: Processing the following image failed"

It is very strange because it only happens sometimes and I've never seen that error before.
I tried by rebooting and also by re downloading everything, updating the nuget packages and many errors appeared. But the "old version" without the nuget updates is working normally except for this can't access the file issue. I'm seeing that about 1/3 of the pictures to be analyzed receive that error, the others are processed without problems.

Hope someone can give me a clue what could be happening.
Thanks
Click to expand...

Hello, I'm experiencing the same problem. Anyone has found a solution yet? This is causing delays in the detection. Thank
 
Greetings.

It's been a while since I tried th AITool and I see that there is a fork with other options besides Telegram. From one of the screenshots of the VorlonCD fork, it looks like it can call a URL such as the trigger URL for a camera in BI (that I already have configured to send me an email when triggered). Is a capture of the detected object to email now possible?

Up until today, I was running the Docker only alternative from neile but broke it due to some attempted updating and am interseted in going back to the original.

Anyone have experience with both the VorlonCD/AITool and the Neile Docker deployments and can shed some light?

Thanks
 
Last edited:
wittaj said:
Here is the API: username: password@IPADDRESS: PORT#/cgi-bin/coaxialControlIO.cgi?action=control&channel=1&info[0].Type=1&info[0].IO=0

Type=1 is for the audio
Type=2 and IO=1 is for the light

No space after the : but if I do not put a space then it shows like this here lol :p

Depending on the camera, you may have to play with the 0 and 1 to figure out which one is which. But the above is the general URL
Click to expand...
Thank-you
 
mcarson75 said:
I am running AI Tool 2.0.319.7666 from 12/27/2020. Everything had been working well and I got the detections working the way that I want. However, this morning I went to check and AI Tool had shut down overnight. I have restarted it and rebooted the machine multiple times. AI Tool runs and the queue fills up with 100 images, but none of them are ever processed. It just stays at "Working".

View attachment 80247

This seems to me to be an issue with Deepstack? But it is running:
View attachment 80248

These are the errors in the log. (Network path is correct now)
View attachment 80249

How can I get this working again? I have re-installed by overwriting the files in the AI Tools directory with the originals.

Edit to add that I have manually tested Deepstack with a POST command and it is running and working properly.
Click to expand...
I'm having the same issue, but I'm using the build from 10/30/2020. I was watching the BI alerts folder to try and figure out why it was saying it can't find the image. It looked like something (probably BI) was renaming the images after they were put in the folder. The AI tool sees the image put in the folder and starts to try and process, but the image is renamed before it can do anything with it. I'm not sure what's going on.
 
Evening all,

Ok so I have managed to install the new ai tool.
Its night time and im getting confidence of 64% on vehicle. in this instance it is correct but it tells me "1x not in confidence range : Truck (62%)" How do i set this range lower so it pics that up.
 
Last edited:
Monty said:
Evening all,

Ok so I have managed to install the new ai tool.
Its night time and im getting confidence of 64% on vehicle. in this instance it is correct but it tells me "1x not in confidence range : Truck (62%)" How do i set this range lower so it pics that up.
Click to expand...
On the cameras tab there is a Confidence limits section where you can change the % of lower and upper limits. What is the lower limit set at on your settings?
 
Brendon06 said:
On the cameras tab there is a Confidence limits section where you can change the % of lower and upper limits. What is the lower limit set at on your settings?
Click to expand...
I have tried 1%-100%, 20%-100%,0-50%. All kinds exactly where you said.
I seem to have a bigger problem now. It will process maybe 1 or 2 images then under over view it just says processing ...jpg and it starts to que other images till 100. Not sure whats wrong
Also getting error no default camera found but I have 2 cameras on
 

Attachments

  • 89fe0b5f8f6717d02cb44b4c9fb2d375.png
    89fe0b5f8f6717d02cb44b4c9fb2d375.png
    37.2 KB · Views: 27
Last edited:
Monty said:
I have tried 1%-100%, 20%-100%,0-50%. All kinds exactly where you said.
I seem to have a bigger problem now. It will process maybe 1 or 2 images then under over view it just says processing ...jpg and it starts to que other images till 100. Not sure whats wrong
Also getting error no default camera found but I have 2 cameras on
Click to expand...
Create a camera called "default" and disable AI detection (assuming you don't want AI for unspecified cameras. This will overcome the default not found error, but unsure what is causing the other messages.
Too many images in your folder? \ AV locking the images? \ Deepstack issue? I run multiple deepstack instances, for redundancy and performance.
 
JamCam said:
I'm having the same issue, but I'm using the build from 10/30/2020. I was watching the BI alerts folder to try and figure out why it was saying it can't find the image. It looked like something (probably BI) was renaming the images after they were put in the folder. The AI tool sees the image put in the folder and starts to try and process, but the image is renamed before it can do anything with it. I'm not sure what's going on.
Click to expand...
Make sure you are not using the BI Alerts Folder.
Recommend setting up a new folder dedicated to store alerts for AI Tools (Blue Iris -> Clips and Archiving" ) then for the Camera select this folder in the Camera -> Record under JPEG Snapshots.
Do you have any examples of a before and after re-name to help determine what is re-naming the files?
 
spammenotinoz said:
Create a camera called "default" and disable AI detection (assuming you don't want AI for unspecified cameras. This will overcome the default not found error, but unsure what is causing the other messages.
Too many images in your folder? \ AV locking the images? \ Deepstack issue? I run multiple deepstack instances, for redundancy and performance.
Click to expand...
I have cleared that folder many times so there is little to no images. I created default camera but still getting the errorf1d12617c206e1824ab0e618922e6492.png
 
What setting would cause alerts to be sent without the subject being captured? I also see that in BI Alerts, that a car is being captured in the latter half of the frame, so I think the alert image being sent is after the car has passed.
 
TrunkMonkey said:
What setting would cause alerts to be sent without the subject being captured? I also see that in BI Alerts, that a car is being captured in the latter half of the frame, so I think the alert image being sent is after the car has passed.
Click to expand...
You could try a longer cool down in bi and shorter in aitools maybe aitools is picking up something but bi has already captured another alert image and used that (being the latest) for alerts
 
Brendon06 said:
You could try a longer cool down in bi and shorter in aitools maybe aitools is picking up something but bi has already captured another alert image and used that (being the latest) for alerts
Click to expand...

I cross checked the alert image against the video in Bi and the image was taken 12 seconds afer the vehicle was fully in frame.

The cooldown in aitool is zero, in BI it's 3s, but changing it to 4s didn't seem to impact it.
 
Can somebody point me to an accurate guide to using the Jetson for the ai device? I have followed the documentation on the Deepstack website and some other places just to get some errors. I can find other people with the same issue, but no real solutions
 
You must log in or register to reply here.
Top Bottom