AI motion detection with BlueIris: an alternate approach

For further assistance with this please open a support request over in the github repo. Make sure to include your complete docker-compose.yml file and the complete output logs from Docker.

Edit: Missed that this was a question about getting AI Tool working with Deepstack installed via my alternate approach. There is no need to use AI Tool when following the approach in this post.
 
Last edited:
neile, thanks for confirming that deepstack is working correctly.

Does anyone know why when i manually trigger the camera using the url, I get the configure alert, but when i walk into the path of any of the cameras I get no alerts? I do have person specified in watch objects in the triggers.json file
 
Does having a space in the path for where /RUN/SECRETS/SETTINGS is mounted make a difference

1597763960314.png
 
What do the logs for the trigger container show? There are tons of logs that will tell you if files were loaded successfully, if there were any errors, whether files are getting picked up and processed, etc.
 
From what I gather in the log there is no blatant error:

2020-08-18T08:07:34-07:00 [Trigger manager] There are 13 images waiting in /aiinput for DoorBellHD detector.

2020-08-18T08:07:34-07:00 [Trigger manager] There are 13 images waiting in /aiinput for DrivewayHD detector.

2020-08-18T08:07:34-07:00 [Trigger manager] There are 13 images waiting in /aiinput for FrontDoorHD detector.

2020-08-18T08:07:34-07:00 [Trigger manager] There are 13 images waiting in /aiinput for GateFrontHD detector.

2020-08-18T08:07:34-07:00 [Trigger manager] There are 13 images waiting in /aiinput for GateRearHD detector.

2020-08-18T08:07:34-07:00 [Pushbullet] Pushbullet is disabled via settings.

2020-08-18T08:07:34-07:00 [Pushover] Pushover is disabled via settings.

2020-08-18T08:07:34-07:00 [Telegram] Telegram is disabled via settings.

2020-08-18T08:07:34-07:00 [Trigger DoorBellHD detector] Listening for new images in /aiinput/DoorBellHD*.jpg

2020-08-18T08:07:34-07:00 [Trigger DrivewayHD detector] Listening for new images in /aiinput/DrivewayHD*.jpg

2020-08-18T08:07:34-07:00 [Trigger FrontDoorHD detector] Listening for new images in /aiinput/FrontDoorHD*.jpg

2020-08-18T08:07:34-07:00 [Trigger GateFrontHD detector] Listening for new images in /aiinput/GateFrontHD*.jpg

2020-08-18T08:07:34-07:00 [Trigger GateRearHD detector] Listening for new images in /aiinput/GateRearHD*.jpg

2020-08-18T08:07:34-07:00 [Main] Watching for changes to /run/secrets/settings

2020-08-18T08:07:34-07:00 [Main] Watching for changes to /run/secrets/triggers

2020-08-18T08:07:34-07:00 [Main] ****

2020-08-18T08:07:34-07:00 [Main] Up and running!

2020-08-18T08:07:34-07:00 [Local storage] Purge complete
 

Attachments

  • 1597764304981.png
    1597764304981.png
    17.8 KB · Views: 346
Yup it's up and running and working fine. Walk in front of your cameras and check to see if you get any logs saying a new image was detected. Check and make sure your aiinput folder is actually getting new images from BlueIris. Make sure your wildcard pattern for the filename matches the file names actually stored in the aiinput folder.
 
No log mentioning new image detected but aiinput folder has images of me in front of camera. Wildcard pattern matches the image filename found in aiinput

------ 8/18/2020 8:36 AM 278467 DoorbellSD.20200818_083653129.jpg

"name": "DoorbellHD",
"watchPattern": "/aiinput/DoorbellSD*.jpg",
"enabled": true,
"threshold": {
"minimum": 10,
"maximum": 90
},
"handlers": {
"webRequest": {
"triggerUris": [""]
},
"mqtt": {
"topic": "aimotion/triggers/DoorbellHD"
},
"telegram": {
"chatIds": [1]
}
},
"watchObjects": ["person"]
 
Is your drop folder a network share that's mounted to your computer running Docker? Or is it a local folder on the same machine as Docker?
 
Well I'm out of ideas. Please open another support request. Make sure to include

  • Your complete docker-compose.yml file and where you have this file stored locally on your machine
  • Your complete settings.json file
  • Your complete triggers.json file
  • The entire log output from the trigger container, from the first line that reports the version number of the system to the last line after you walk past a camera and don't see any detection
 
I found the problem and it's my own fault with editing the triggers.json file.The log is busily analyzing images now. Thanks again for your help.
 
  • Love
Reactions: neile
I found the problem and it's my own fault with editing the triggers.json file.The log is busily analyzing images now. Thanks again for your help.

What was the problem? Maybe there's a way I can add some logs to detect that situation for others.
 
There's no log to fix lack of attention and repeated stupid mistakes. I was so burnt out on trying to figure it out I repeatedly overlooked the same mistakes you instructed me to check over and over.

One suggestion I have is to build out an easier way to edit, check for consistency and create the triggers.json and docker-compose.yml so that simple mistakes can be avoided that also allows for adding cameras in mass. I occasionally use a tool to create Microsoft unattend.xml files for unattended deployment of Windows OS on computers that allows you to input or select from some parameters then outputs your options into a properly formatted file. Windows Answer File Generator
 
There's no log to fix lack of attention and repeated stupid mistakes. I was so burnt out on trying to figure it out I repeatedly overlooked the same mistakes you instructed me to check over and over.

One suggestion I have is to build out an easier way to edit, check for consistency and create the triggers.json and docker-compose.yml so that simple mistakes can be avoided that also allows for adding cameras in mass. I occasionally use a tool to create Microsoft unattend.xml files for unattended deployment of Windows OS on computers that allows you to input or select from some parameters then outputs your options into a properly formatted file. Windows Answer File Generator

Did you use Visual Studio Code to edit the files as mentioned in the documentation? It will provide full Intellisense auto-complete and validation of the files as you type.
 
Did you use Visual Studio Code to edit the files as mentioned in the documentation? It will provide full Intellisense auto-complete and validation of the files as you type.

I did initially, but I switched back to Notepad++ as I was more used to the layout of that than Visual Studio Code.
 
Hi,

Do I need a GPU for this version of deepstack? Does it benefit from a GPU and if so what is recommended? I have ordered a dedicated Dell OptiPlex 5070 SFF Desktop i7-8700 with 16GB RAM 256GB SSD W10 PRO for BI. I intended to run Deepstack on this machine via a docker. Is 16GB of RAM enough for both?
 
Hi,

Do I need a GPU for this version of deepstack? Does it benefit from a GPU and if so what is recommended? I have ordered a dedicated Dell OptiPlex 5070 SFF Desktop i7-8700 with 16GB RAM 256GB SSD W10 PRO for BI. I intended to run Deepstack on this machine via a docker. Is 16GB of RAM enough for both?

You do not need a GPU. 16GB of RAM is plenty.
 
Hi,

Do I need a GPU for this version of deepstack? Does it benefit from a GPU and if so what is recommended? I have ordered a dedicated Dell OptiPlex 5070 SFF Desktop i7-8700 with 16GB RAM 256GB SSD W10 PRO for BI. I intended to run Deepstack on this machine via a docker. Is 16GB of RAM enough for both?

I’m running on an i5 with 12gb of ram. I think you could use the gpu, but you’ll have to switch to that version of deep stack in the .yml file before you run the docker up command. (It’s easier to use cpu only)