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

I'm sorry you seldomly hear things from my end and thankfull VorlonCD is doing such a great job! Today, I got myself a Jetson Nano running too and I'm sharing the install process to help everyone who hasn't yet become friends with linux terminal entirely ;).

Guide to set up Deepstack on the Nvidia Jetson Nano 2GB

Thanks for the tutorial! I was able to set it up using your instructions. I'm using the 4GB Jetson and it took 10 seconds for the first analysis and 700 to 800 ms for the rest. Are you also getting similar speeds?
 
Hello!

Very happy to have this resource. I've been working on it for a few days and can't seem to get past this point. Seems Python is giving an error at startup. I can't find that file it's looking for either. Did I miss a setup step for Python?

Thank you

Debug: DeepStack>> Init: File "../init.py", line 9, in <module>
DeepStack>> C:\DeepStack\interpreter\python.exe: can't open file '../intelligence.py': [Errno 2] No such file or directory
Im having the exact same problem, so please let me know if you have figured it out.
 
I'm sorry you seldomly hear things from my end and thankfull VorlonCD is doing such a great job! Today, I got myself a Jetson Nano running too and I'm sharing the install process to help everyone who hasn't yet become friends with linux terminal entirely ;).

Guide to set up Deepstack on the Nvidia Jetson Nano 2GB

1. Getting Started with Jetson Nano 2GB Developer Kit
2. Getting Started with Jetson Nano 2GB Developer Kit
-> install in headless mode

Thank for this excellent guide to setting up the Jetson Nano. I purchased one to take some of the load off my BI server PC but it works so well that I have removed deepstack from my BI server completely! I have 15 cameras of mixed makes most of which are 1080. JPG quality for AITool is set to 90%. Typically I'm getting deepstack timings of 400ms in "high" mode. (It's connected via ethernet)

Hopefully, the inclusion of currently limited support in BI for deepstack won't put VorlonCD off further development. Having the ability to tailor our setups with the huge variety of options and settings will never be included in an program like BI and if it was it would probably be an extra subscription!

Thank you GentlePumpkin and VorlonCD for a brilliant program.
 
Hello!

Very happy to have this resource. I've been working on it for a few days and can't seem to get past this point. Seems Python is giving an error at startup. I can't find that file it's looking for either. Did I miss a setup step for Python?

Thank you

Debug: DeepStack>> Init: File "../init.py", line 9, in <module>
DeepStack>> C:\DeepStack\interpreter\python.exe: can't open file '../intelligence.py': [Errno 2] No such file or directory
Im having the exact same problem, so please let me know if you have figured it out.

I have the same problem. For a workaround you can run DeepStack prior to opening AITool.

deepstack.exe --VISION-SCENE=False --VISION-DETECTION=True --VISION-FACE=False --PORT 81
 
  • Like
Reactions: lordicon232
I ended up fixing the python error by delete the file sin this cache folder. I did on a whim. Seemed like my error message was pulling from a dead/corrupt DB.

I cleared the files that were inside: C:\DeepStack\windows_packages\pycache
 
Anyone know how to set the docker time in Docker Windows Desktop? It seem to be showing UTC not UTC+2 that the computer is running.
 

Attachments

  • Workbench-2.png
    Workbench-2.png
    7.9 KB · Views: 28
  • Workbench-3.png
    Workbench-3.png
    1.5 MB · Views: 29
hi, I am using the AiTool for a long time now,
but after update with version AITool-VorlonCD-v2.0.1014
the Trigger URLs in Actions do not work correct anymore,
if you enter multiple URLs - sometimes only one, sometimes none trigger.
Can you please check this?
 
First of all, this is a great initiative, I just found out about it and had to set my BI up to do this, and after a few hours of reconfiguration, things are working great, the people/auto detection is so much better now.

My setup:
BI version 4 with 8x outdoor IP camera running on Windows Server
Docker DeepStackAI running on my Synology NAS
Vorlon-2.0.1068 running on Windows Server

I have a logging issue in AI Tool (Vorlon-2.0.1068) where I constantly get the following error, once every couple of seconds, everything is working fine otherwise. I am not sure why AI Tool constantly tries to set Registry.
Insufficient system resources exist to complete the requested service.; [IOException] (In Global.SaveRegSetting(String name, Object value, String SubKey):1715 > Win32.RegistryKey.SetValue(String name, Object value, RegistryValueKind valueKind) > Win32.RegistryKey.Win32Error(Int32 errorCode, String str))
 
hi, I am using the AiTool for a long time now,
but after update with version AITool-VorlonCD-v2.0.1014
the Trigger URLs in Actions do not work correct anymore,
if you enter multiple URLs - sometimes only one, sometimes none trigger.
Can you please check this?

i am had the same issue with v2.0.1015 It would send alerts through Telegram, but BI would not alert sometimes. Other times it would. I went back to 2.0.527.7681 built on 1/11/2021 and i think it is working again. I think it was my settings after upgrade, but not sure.
 
I had similar woes then learned the URL was pointing to wrong IP address. I checked the variables and found the BlueIrisURL was for my NIC that was only inside camera subnet. In BlueIris, Settings > Web Server, internal, I changed the IP to NIC that was visible to AiTools and that got the alerts working as AiTools could now find BI.
 
Im trying to upgrade my notifications from home assistant, the issue being cropped images in the notification.
standard processed image is 640x480 (sent from AItool via MQTT to HA), the notification on my phone is 640x240 i think.. if the detection is right at the top or right down the bottom i cant see it..

So to try and mitigate this ive setup 2 camera proxies for each camera where there is a high or low crop and using a template based on the position of the detection I decide which proxy to send on the notification, its a very ugly setup

Is there a simpler way to achieve this? or would it be possible to have an option to set the MQTT image size and have the detection centred on the image?
 
From appearances, and as a long time lurker, this topic/thread has slowed down drastically since BI has added native Deepstack support. Personally I have not looked into the native support capabilities as BI runs on 1 system and the AI / Deepstack on another etc.

Is continued development of the AI Tool being planned in light of the above?
 
From appearances, and as a long time lurker, this topic/thread has slowed down drastically since BI has added native Deepstack support. Personally I have not looked into the native support capabilities as BI runs on 1 system and the AI / Deepstack on another etc.

Is continued development of the AI Tool being planned in light of the above?

I'm pretty sure it is as it offers so much more.
 
From appearances, and as a long time lurker, this topic/thread has slowed down drastically since BI has added native Deepstack support. Personally I have not looked into the native support capabilities as BI runs on 1 system and the AI / Deepstack on another etc.

Is continued development of the AI Tool being planned in light of the above?
I still use AITool as there is a lot more granular control over things like percentage size of detected object, percentage of confidence per object (in the newer versions).
I have cameras up on second story due to the way the place is laid out and AITools can be set to pick up objects that the native BI integration misses.

Even if development on AITools stopped right now and didn't proceed further than it is currently at, you still have a lot more granular control than the native integration has. and it would continue to be usable for a long time.
 
I still use AITool as there is a lot more granular control over things like percentage size of detected object, percentage of confidence per object (in the newer versions).
I have cameras up on second story due to the way the place is laid out and AITools can be set to pick up objects that the native BI integration misses.

Even if development on AITools stopped right now and didn't proceed further than it is currently at, you still have a lot more granular control than the native integration has. and it would continue to be usable for a long time.
Dealbreaker for me is the possibility to run it on a different computer. I think deepstack on the same machine with lots of cameras will eat alot of CPU...
 
  • Like
Reactions: originalcom
Dealbreaker for me is the possibility to run it on a different computer. I think deepstack on the same machine with lots of cameras will eat alot of CPU...
I run deepstack with lots of cameras on the same machine with no problem. I do use the gpu version. I'd still use AITool though.
 
I run deepstack with lots of cameras on the same machine with no problem. I do use the gpu version. I'd still use AITool though.
Well, it all depends on the machine I guess... Its much easier to get away with cheaper machines if I split it up, which means I can get better cameras ;)