Hi GentlePumpkin,
Yes you are correct. For BI v5 'Capture an alert list image' option needs to be enabled.
Yes you are correct. For BI v5 'Capture an alert list image' option needs to be enabled.
[tool] [tutorial] Free AI Person Detection for Blue Iris
- Thread starter GentlePumpkin
- Start date
- Sep 4, 2017
- 194
- 325
Add me to the list of successful installs. I have it up and running on my BI5 install running on Win10Pro. Only have it setup for one camera at the moment but so far so good.
I did follow the tutorial but had some questions. During the part (1.2) to get deepstack server running, it says to select Detection API but doesn't say if I should unselect Face API. Also, you say to input your API key. Is that the activation key? I can't find an API key and the server says the API keys are optional. I did leave those fields blank and it's working.
During step 2.4 when you disable all unnecessary stuff, I assume that also means disable any zones and object detection stuff? What about the basic motion settings? Do I move the min object size and min contrast all the way to the left?
Thanks for such great work on this. I really appreciate all your work and efforts.
I did follow the tutorial but had some questions. During the part (1.2) to get deepstack server running, it says to select Detection API but doesn't say if I should unselect Face API. Also, you say to input your API key. Is that the activation key? I can't find an API key and the server says the API keys are optional. I did leave those fields blank and it's working.
During step 2.4 when you disable all unnecessary stuff, I assume that also means disable any zones and object detection stuff? What about the basic motion settings? Do I move the min object size and min contrast all the way to the left?
Thanks for such great work on this. I really appreciate all your work and efforts.
- Sep 4, 2017
- 194
- 325
You can leave it activated, I usually disable it because it might cause that DQAI requires more performance.
The DQAI activation key is meant. Were you able to use DQAI without ever entering the activation key? (Because that would be really cool
)In the camera that the AiTool triggers, I disabled motion detection entirely.
Thanks for your positive feedback
My system has been running flawlessly for some time but last night I did some housekeeping and installed a new disk drive. I was copying all the files from a 500Gb hard drive to a 6T drive and I kept getting the errors you mentioned. Both drives didn't have any of the image files on and neither are the location where they are saved. The CPU was only running at about 40% so not maxed out and I have plenty of available RAM.. As soon as the copy was finished then no more errors.
sansnil
n3wb
- Oct 7, 2016
- 12
- 3
That's interesting thanks for sharing. I think it also correlates the theory that the computer OS must be 'holding' on to the image files a bit longer than the current version of AI Tool is looking to grab them (500 ms as I recall from GentlePumpkin's post). I'm interested to see if allowing a longer time to pickup the file gets rid of the errors.
sansnil
n3wb
- Oct 7, 2016
- 12
- 3
I like that idea - a setting for changing the default to a custom wait setting would be great as it could optimized for each person's setup.
traceray
n3wb
- Mar 9, 2017
- 1
- 0
Great Job Gentle Pumpkin! It's an excellent tool. I am currently using deprecated version 0.6. Do you think it's possible to share the source code of that version? I would like to make it detect multiple people, not only one. The current versions has multiple object and person detection but they dont crop the detections to an output directory.
Thanks in advance
Thanks in advance
Help, this is not working for me. My BI uses port 81, should the server for the AI be on the same port as I have put it on 80?
Thanks in advance.
Thanks in advance.
Sorry probably should have said, this is what I'm getting in the log:
14.10.2019, 22:48:54]: ERROR: Processing the image E:\AI/aigarden.20191014_223104662.jpg failed. Can't reach DeepQuestAI Server at localhost:80.
[14.10.2019, 22:50:34]: ERROR: Processing the image E:\AI/aigarden.20191014_223105664.jpg failed. Can't reach DeepQuestAI Server at localhost:80.
[14.10.2019, 22:52:14]: ERROR: Processing the image E:\AI/aigarden.20191014_223106665.jpg failed. Can't reach DeepQuestAI Server at localhost:80.
Try starting the DeepQuestAI server on a different port. Since your BI is using port 81, and port 80 isn't working for the AI, try port 82. I've seen stock installs of Win10 use port 80. I discovered that when my BI wouldn't run on port 80. I don't remember the service it was, but I had to disable said service in order to use port 80 with BI. You do want both BI and AI on different ports. I use 80 for BI and 81 for AI.
Thanks for your help, I have tried this with no luck. I can access the deep stack server by typing localhost
ort and can accesses it from other computers on the same network so im not quite sure what's going wrong 
Fantastic app. Thank you for releasing this!
I have setup this and it is working almost perfectly except for the following error when trying to call a url:
[16.11.2019, 20:40:54]: trigger url: [16.11.2019, 20:40:54]: ERROR: Could not trigger URL 'x', please check if 'x' is correct and reachable.
[16.11.2019, 20:40:54]: -> Trigger URL called.
[16.11.2019, 20:40:54]: -> Trigger URLs called.
Is there something in my URL that isnt supported?
Update: Used a URL shortener to strip out the extra stuff. Not ideal but it works now!
Second question: I see in the release notes I can notify based on the type of object detected - what is the syntax for the URL? Sorry if I am blind...
I have setup this and it is working almost perfectly except for the following error when trying to call a url:
[16.11.2019, 20:40:54]: trigger url: [16.11.2019, 20:40:54]: ERROR: Could not trigger URL 'x', please check if 'x' is correct and reachable.
[16.11.2019, 20:40:54]: -> Trigger URL called.
[16.11.2019, 20:40:54]: -> Trigger URLs called.
Is there something in my URL that isnt supported?
Update: Used a URL shortener to strip out the extra stuff. Not ideal but it works now!
Second question: I see in the release notes I can notify based on the type of object detected - what is the syntax for the URL? Sorry if I am blind...
Last edited:
From here here
So dont think about it too much. Go install, configure it and use it! You will thank yourself for doing that soon after
I haven't tried Sentry, but this has been working well for me.
I am monitoring a camera that is pointed towards my front door from above and to the side. I have Blue Iris set to save JPGs into the folder for analysis only when motion occurs in the area near the door and I am using a mask in the AI Tool to only look for people in a small area near the door. I am currently saving one image every 2 seconds when motion is detected there. Sometimes the timing of the images places the person either outside the mask or behind a post that obscures them from the cameras view so I see that the image was analyzed but the AI doesn't see the person. I'm sure I could fine tune this by adjusting the mask area and the time between images saved, but it hasn't been that frequent so I haven't gotten around to it yet.
Probably the most confusing part of the setup is getting a bot set up on Telegram for the alerts, but there are decent tutorials online to help with that.
The Telegram alerts seem to be pretty quick and reliable. The alert goes to both my phone and my wife's phone (with a doorbell sound for the alert) and we almost always receive the alerts within a second of each other. Every once in a while Telegram seems to have a delay in sending messages, but it's pretty rare. Also I've noticed that after starting Deepstack the first image analysis takes a long time (20 seconds or more) but after that it typically takes Deepstack 1.2 seconds to analyze the image on my system (i5 4670 with Blue Iris and several other apps running on the same box). From the time a person is in view to the alert on the phone is usually just a few seconds total. I will probably drop the time between JPGs in Blue Iris to 1.5 seconds so it will typically write images almost as fast as Deepstack can process them.
I had a few missed alerts when the camera was in night mode (black & white) and when I reviewed the images it looked like the contrast of the person against the background was really low, so I switched the camera a few days ago to stay in day mode all the time. There is enough light near the front door that I think this will work better.
Overall I'm very happy with the AI Tool.
I am monitoring a camera that is pointed towards my front door from above and to the side. I have Blue Iris set to save JPGs into the folder for analysis only when motion occurs in the area near the door and I am using a mask in the AI Tool to only look for people in a small area near the door. I am currently saving one image every 2 seconds when motion is detected there. Sometimes the timing of the images places the person either outside the mask or behind a post that obscures them from the cameras view so I see that the image was analyzed but the AI doesn't see the person. I'm sure I could fine tune this by adjusting the mask area and the time between images saved, but it hasn't been that frequent so I haven't gotten around to it yet.
Probably the most confusing part of the setup is getting a bot set up on Telegram for the alerts, but there are decent tutorials online to help with that.
The Telegram alerts seem to be pretty quick and reliable. The alert goes to both my phone and my wife's phone (with a doorbell sound for the alert) and we almost always receive the alerts within a second of each other. Every once in a while Telegram seems to have a delay in sending messages, but it's pretty rare. Also I've noticed that after starting Deepstack the first image analysis takes a long time (20 seconds or more) but after that it typically takes Deepstack 1.2 seconds to analyze the image on my system (i5 4670 with Blue Iris and several other apps running on the same box). From the time a person is in view to the alert on the phone is usually just a few seconds total. I will probably drop the time between JPGs in Blue Iris to 1.5 seconds so it will typically write images almost as fast as Deepstack can process them.
I had a few missed alerts when the camera was in night mode (black & white) and when I reviewed the images it looked like the contrast of the person against the background was really low, so I switched the camera a few days ago to stay in day mode all the time. There is enough light near the front door that I think this will work better.
Overall I'm very happy with the AI Tool.
Spackler
n3wb
- Apr 27, 2018
- 28
- 4
Thank you for the responses. I am going to get it setup now. Any suggestions, lessons learned or things to look out for while I am setting up?
Disable unneccessary stuff - What all should be disabled?
My Triggers are not working either. I am getting prompted to enter username and password even though I have it in the address.
ERROR: Could not trigger URL 'x', please check if 'x' is correct and reachable.
Can I also expect not to see anymore short videos in my alerts?
How can I add them back as I use the BI app on my phone so that I can see the alerts and video to what the AI has alerted me to.
Thanks
Disable unneccessary stuff - What all should be disabled?
My Triggers are not working either. I am getting prompted to enter username and password even though I have it in the address.
ERROR: Could not trigger URL 'x', please check if 'x' is correct and reachable.
Can I also expect not to see anymore short videos in my alerts?
How can I add them back as I use the BI app on my phone so that I can see the alerts and video to what the AI has alerted me to.
Thanks
Last edited:
For triggers - I dont have any issues with this as I have added an exemption for the BI server itself so I will not require a password (BI Settings>>Web Server>>Advanced>> Limit IP Addresses - enter your BI ip address in there using this format ^192.168.100.1, then press OK.
In AI Tool this is the format I use for triggering - (change the IP address and camera short name as per your settings of course)
You should be seeing alerts in BI (and BI client). I believe you dont see them as it might be a configuration issue. Below are my settings for the original camera and the cloned camera. Use these as a starting point and customize as per your requirements.
Also I believe Disable unnecessary stuff is about the settings on the cameras - only enable what is required. For example in the cloned camera no need for audio or record, etc so in my example these settings are only enabled in the original camera settings.
Original Camera
Cloned Camera
Hope this helps.
Last edited:
Spackler
n3wb
- Apr 27, 2018
- 28
- 4
Thank you the triggers seems to not be erroring now. While I was configuring I got two alerts in the BI app, but nothing since. I setup my cameras original and clone just like yours. I got alerts np till setting up the ai. Sure I am missing something simple. Any ideas to get those back would be great. I really want to be able to see the alert video once the ai notifies me via Telegram. Thanks again for all the help.
The AI software seems to not be triggering for it to log in BI. I had my webserver set to 510 but I changed it to 80 and still no trigger.
Last edited: