CodeProject.ai with License Plate Reader module - HowTo?

is there a more GUI way of seeing the plates caught other than shifting manually through the logs?

Use this utility I referenced in your other post.

 
  • Like
Reactions: MikeLud1
Use this utility I referenced in your other post.


I have looked at that over the past couple of days and got totally lost. the post you referenced seems to be mid way through several other pages of threads, etc... do you know of anywhere that there are steps to follow, ie, step 1) download/install this... step 2) change this config, step 3) etc... ?
 
I have looked at that over the past couple of days and got totally lost. the post you referenced seems to be mid way through several other pages of threads, etc... do you know of anywhere that there are steps to follow, ie, step 1) download/install this... step 2) change this config, step 3) etc... ?

Um I clicked that link and it opens on post #1 of that thread on NodeRed complete with the steps?

 
@MikeLud1 - Any chance you are around today to help me figure out how to get the plates to come kind of graphic interface? ie, PlateRecognizer or something better?
 
I'm SOOOOO annoyed right now... we just had the vehicle that the neighborhood suspects of all the thefts in the neighborhood drive by my cameras - and for whatever reason, they didn't trigger to catch the video and plates. Like they didn't trigger at all from the rear of the vehicle side. They did trigger from the headlight side, but of course no plate on that side. I have my zones set up to basically cover the entire road in each direction. I have my motion set to trigger with a min. obj size of 100 and min contrast of 12 (the lowest possible settings) and I have my object travels set to 25, also tried 10... which is crazy low... and yet, on the back side of the vehicle at night time it's NOT triggering capture. It was triggering correctly prior to making the changes I made to get LPR to work... am I missing a setting somewhere or is the LPR settings/solution somehow making the trigger less sensitive? Any thoughts @MikeLud1 or anyone else? Thanks!

Sensor.png

Zone.png

objDetection.png

trigger.png
 
You cannot use Edge Vector at night, especially for plates.

I mentioned previously, uncheck the object moves X number of pixels and uncheck the reset at X% - those are only used to eliminate false triggers and you have the opposite going on. NONE of my cameras use those settings.

If you are not recording continuously, you should also do that in the event motion doesn't trigger an alert.
 
  • Like
Reactions: CCTVCam and Mike A.
Since you are using Code Project to weed out the false triggers, I would use simple algorithm for both day and night and let Code Project decide if it meets the AI criteria.
 
  • Love
  • Like
Reactions: CCTVCam and actran
If you are not recording continuously, you should also do that in the event motion doesn't trigger an alert.

This. Any important cams you should run continuous in some form. Substreams at least. Too many times things can be missed otherwise and storage is cheap.
 
  • Like
  • Love
Reactions: CCTVCam and actran
This. Any important cams you should run continuous in some form. Substreams at least. Too many times things can be missed otherwise and storage is cheap.

I put the two license cams on continuous. I have them on a 10TB drive... How many days do you think I'll get out of that?
 
I put the two license cams on continuous. I have them on a 10TB drive... How many days do you think I'll get out of that?

Depends on the resolution and bitrate and what else is going to it:


And then using substream will get you more storage based on how much it triggers to mainstream.
 
  • Like
Reactions: Mike A.
Recording a D1 substream with a decent bitrate won't take up much space but will also capture those that don't trigger:

 
  • Like
Reactions: Mike A.
Does anyone have any idea where these images are stored? I'm trying to fine tune my trigger points and being able to see what snapshots it is looking at to analyze the image for a plate would help. My license plate reader is functioning, but is very random on whether or not it finds a plate to analyze. The plates are VERY clear in the video's.

ai.png

status.png
 
Does anyone have any idea where these images are stored? I'm trying to fine tune my trigger points and being able to see what snapshots it is looking at to analyze the image for a plate would help. My license plate reader is functioning, but is very random on whether or not it finds a plate to analyze. The plates are VERY clear in the video's.

View attachment 169966

View attachment 169967
Right click on the alert and click on Open Containing Folder. Also you can hold down the CTRL key and double left click on the alert and the last screenshot will popup.

1691887063603.png
1691887271337.png
 
  • Like
Reactions: actran
Right click on the alert and click on Open Containing Folder. Also you can hold down the CTRL key and double left click on the alert and the last screenshot will popup.

View attachment 169970
View attachment 169971

I don't have anything like that in my clips on the left, definitely not any zoomed in plate images... I just have the 5-20 second video clips over there. nothing when I click the alerts lightning bolt.

bi2.png

ai2.png
 
Last edited: