Possibly... Did you try sending yourself an email alert and including &JSON in the message? It should have all the details from the call to the ALPR module. If it doesn't, then something is messed up. Can you add an email and post the JSON contents?
Make sure you have the correct call. Don't put the &JSON in quotes after the ai_dump parameter.
{ "ai_dump":&JSON, "Image":"&ALERT_JPEG", "camera":"&CAM", "ALERT_PATH": "&ALERT_PATH", "ALERT_CLIP": "&ALERT_CLIP", "timestamp":"&ALERT_TIME" }
It looks like your &JSON is empty.
If still having...
The license-plate custom model is to find a rectangular region consistent with a license plate. It does not, on it's own, figure out the letters and digits on the plate.
Putting ALPR DB aside, are you getting plates read in BI correctly first?
OODPCVS = Overview, Outline, Discern, Perceive, Characterize, Validate, Scrutinize.
I don't know why there need to be 7 levels, especially since we can just say "Enhance" and get the Scrutinize level of video. ;)
Any chance this will also allow the adding of a YOLO11 object detection module? I know you were trying to get Chris Maunder to add it to the CPAI codebase before it got stalled. Selfishly I'd like to test my delivery models on YOLO11.
There isn't much reason to have both New and Storage on the same machine otherwise you are spending all your time copying and deleting.
Most people set it up so their large HDD is New and then delete when that is about 90% full. Storage folder is then unnecessary.
As for your CPU usage, that...
I'd read through the following thread and see if anything is relevant to you. In particular, since CPAI moved off of the CodeProject servers, some links for updates in 2.6.5 won't work. But if it is working for you, consider "if it ain't broke..."...
Are you having issues with 2.9.5? I'm not, so I'm sticking with that version.
Would I like to see active updates? Yes, but I'm content with a working system, for now.
WIll this be added to a CPAI build, or is Chris Maunder no longer making changes so we have to do our own patches? I would love to see a loadable YOLO11 module while we're at it. I know you (@MikeLud1) were working with Chris to get one added, but I haven't seen any traction on that request in...
I reread @MikeLud1 's post and he does indeed say he is thinking of adding color, make and model. So ignore what I just said about that. Impressive if he can do that.
But, that's yet another piece of data that we could use to match misread plates. If a Red Ford Bronco with plates AOT123 is seen...
@MikeLud1 can correct me, but I think the screenshot you saw where it was detecting car details like the make and model was coming from PlateRecognizer.
He was showing an example where the plate was at angle where his updated YOLO11 model was doing better than PlateRecognizer. But I don't think...
I don't feel it's "principally incorrect" to use additional information from the database.
The recognition can always be changed to a better model, a more accurate model, or even a different service (e.g. PlateRecognizer). That change should be independent of what additional steps the database...
I'm using PlateRecognizer for OCR and it does a pretty decent job of character recognition. But even that can have issues with a slight change in lighting, an obstruction, change in angle of travel, front vs. rear, etc. Expecting the OCR to be 100% correct is unrealistic.
The big benefit I see...
Latest nextdoor post:
" URGENT NEIGHBORHOOD ALERT: Thieves Using Wi-Fi Jammers!
Hey Neighbors, I wanted to share a critical warning that everyone with security cameras needs to know about.
Our daughter's bicycle was recently stolen and our cameras didn't record anything. We did some reading...
AI isn't magic. If I look quickly at those pixels I see a person-like figure rather than a typical 4 legged equine.
I expect AI to make mistakes but that's true whether it's running on Deepstack, CPAI or Blue Onyx. The recognition is only as good as the model you are using and the specific...
Could it be related to the case change of "Index" to "index" that was recently made by Ken?
https://ipcamtalk.com/threads/5-9-9-january-17-2025-more-pre-6-0-feature-releases.80127/post-949964
As I found, that doesn't seem to work until a user is manually logged in. So I have to RDP to the BI machine which logs in as a user, and then starts the docker engine.
Again if anyone knows of a way to do this automatically in Windows (short of disabling user credentials) several of us would...
I asked about running docker as a service in Windows awhile ago, but wasn't able to find a solution. If you figure something out, let me know.
https://ipcamtalk.com/threads/full-alpr-database-system-for-blue-iris.79082/post-899295
So @algertc, what's in the next update? Are you working on the AI integration? Or auto correction of plates? I'd also like to put in a reminder request for a method of creating rules for notifications, etc.
If anyone else agrees, please add an upvote to this roadmap entry...
For direction of travel, in my case I have two zones (left and right) with motion triggered from A>B. Then I cloned the camera and changed it to B>A. So I end up with two virtual cameras with cars traveling one way detected on the first and in the other on the second.
A camera that only can zoom is always centered on the middle of the view. In your case you want that to be the bus stop which is up and to the left.
If you zoomed the view pictured, you'd be looking at your immediate neighbor's mailbox and driveway.
If you moved the camera so it was centered...
Don't forget the awesome ALPR DB system that @algertc has created. If people are recognizing plates but not yet logging them into a database, this is definitely something to consider.
https://ipcamtalk.com/threads/full-alpr-database-system-for-blue-iris.79082/
The problem I see is that area is not in the center of your current view. If all the camera can do is zoom, then you'd need the zoomed in and zoomed out views to have the same center.
So a PTZ instead?
From Nextdoor:
"Last night someone robbed a home on [redacted] around 9:03pm. There are plenty of good quality cameras on this court, the low life loser pieces of S, WILL get caught, that’s for certain.
They had to scale a fence and knew what they were going after in the home. It was more than...
I'm using Plate Recognizer for my detections and it is not infallible either. There are times with changing lighting conditions, physical obstructions, etc. where it gets a "close" but slightly wrong plate. Whether it be a database of corrections that have been made previously that it can auto...