Can you run Deepstackand and Plate Recognizer at the same time? I am no longer getting tag numbers on my alerts now that I have Deepstack running.
Thank you,
Rick
Thank you,
Rick
Question
- Thread starter Skinny1
- Start date
Yes, but you need to re-enter your Plate Recognizer key into it at the global level in AI and then select within the camera setting the check box for Plate Recognizer.
I do not see a global level box in Al. I see three boxes, The top one is where I pasted my key too. The second box there is a . The third box is USA Kansas.
Yes, that is the global level when you click on the gear icon and then AI.
Now go to the camera settings for the camera that you are using for Plate Recognizer and then go to the Trigger tab and then the When Triggered area select Artificial Intelligence to go to the next screen and at the bottom is the check box for Plate Recognizer.
Now go to the camera settings for the camera that you are using for Plate Recognizer and then go to the Trigger tab and then the When Triggered area select Artificial Intelligence to go to the next screen and at the bottom is the check box for Plate Recognizer.
Looks like I have it all set up correctly. I'll keep messing with it. Looks like 08/2020 they updated the server software to either do live streaming or snapshot. I picked live streaming. Maybe that is my problem. I should have picked a snapshot instead.
One last question, are you running Plate Recognizer on the Cloud API or On-Premise SDK Port? Thank you for all your help, I appreciate it.
This is weird. Above my key in Ai it is showing "Calls for 2021/4: 1519/2500". So I know it is working but it is not showing the tag numbers on alerts videos or JPG pictures.
What computer do you have and what is the CPU load?
What I think is happening is the Deepstack AI is spiking your CPU just enough that the timing is now off on your alert pics sent to Plate Recognizer or the ordering sequence of BI is causing it. When I was first trying Plate Recognizer, I created a clone so I could test it independent of my already tuned in camera. What I experienced with a cloned camera and not changing a single thing showed a slight delay. I chalk it up to the computer or camera can't keep up and results in a tiny lag or it could be BI or some combination. Maybe it processes the first camera and then the clone because the vehicle will always be in just a slightly different location in the snapshot between the camera and clone. And occasionally I also get where one gets the trigger event and the other camera didn't trigger for the same event. Now mine is for a plate reader that is zoomed in tight, so a fraction delay and a fast moving car and it can miss it. At least that is my experience. I had to adjust my motion settings when I went to this.
You may need to back down how many cameras you are using for Deepstack and look at changing your motion settings to see if you can line it up.
What do the images look like to the left in the alert clips - is the front or back of the car with the plate in the frame or are they off now? I suspect they are now off. Plate Recognizer uses the leading alert clip image, so dialing that in is key and may have to be adjusted over time as we add more bells and whistles to our systems.
What I think is happening is the Deepstack AI is spiking your CPU just enough that the timing is now off on your alert pics sent to Plate Recognizer or the ordering sequence of BI is causing it. When I was first trying Plate Recognizer, I created a clone so I could test it independent of my already tuned in camera. What I experienced with a cloned camera and not changing a single thing showed a slight delay. I chalk it up to the computer or camera can't keep up and results in a tiny lag or it could be BI or some combination. Maybe it processes the first camera and then the clone because the vehicle will always be in just a slightly different location in the snapshot between the camera and clone. And occasionally I also get where one gets the trigger event and the other camera didn't trigger for the same event. Now mine is for a plate reader that is zoomed in tight, so a fraction delay and a fast moving car and it can miss it. At least that is my experience. I had to adjust my motion settings when I went to this.
You may need to back down how many cameras you are using for Deepstack and look at changing your motion settings to see if you can line it up.
What do the images look like to the left in the alert clips - is the front or back of the car with the plate in the frame or are they off now? I suspect they are now off. Plate Recognizer uses the leading alert clip image, so dialing that in is key and may have to be adjusted over time as we add more bells and whistles to our systems.
Last edited:
I am running a Server setup with Blue Iris running an Intel Core I7-6700 @ 3.40 GHz with 16 GB Ram. I have 20 camera's of which I switched to only 11 the running Deepstack and only two running Plate Recognizer. So we'll see if that helps.
Yeah, so that is telling me that it is processing the Deepstack first and by the time that process happens, the vehicle is out of the picture.
I have tuned mine to only people on my property so it isn't sending vehicles in the street to DeepStack. So maybe try to further refine which cameras you want sending to Deepstack. At a minimum not your cameras being used for Plate Recognizer. But then try not to send any cameras that would be sending to Deepstack the same time a camera is sending to Plate Recognizer.
I have tuned mine to only people on my property so it isn't sending vehicles in the street to DeepStack. So maybe try to further refine which cameras you want sending to Deepstack. At a minimum not your cameras being used for Plate Recognizer. But then try not to send any cameras that would be sending to Deepstack the same time a camera is sending to Plate Recognizer.
looney2ns
IPCT Contributor