BI & Plate Recognizer
- Thread starter OICU2
- Start date
I haven't quite made it over to v5 just yet. Hopefully this weekend when there's a better downtime for the system. Everything is backed up and ready to go.
That being said, is Plate Recognizer the "built in" LPR in v5? I thought it was a seperate entity, at an additional cost? Right now I'm using OpenALPR with two cameras. I was looking forward to having everything "all in one" with BI.
That being said, is Plate Recognizer the "built in" LPR in v5? I thought it was a seperate entity, at an additional cost? Right now I'm using OpenALPR with two cameras. I was looking forward to having everything "all in one" with BI.
What does that log file look like? Can you send a screenshot? Wanting to know if it is in a format that could easily go into excel and you can do search and sort by plate?
Thanks - looks like with minimal cleanup, one could do that.
I am on the fence between using the one integrated with BI or going the OpenALPR route and go thru the steps people have documented here on how to automate it on that format.
I am on the fence between using the one integrated with BI or going the OpenALPR route and go thru the steps people have documented here on how to automate it on that format.
Let us know what you decide and how it works out.
So the other day I went ahead and tried plate recognizer. For about an hour or so it wasn't sending any pics, but then all of a sudden it started LOL.
Wow a lot more cars come past than I thought. I am already close to the halfway point of the 2,500 lookups
The simplicity of it is easy. And it is doing a fairly accurate job, even on plates that are total crap, but I know what they are. I was surprised and it gets them fast too.
It is simple enough to take the BI log and import into excel and quickly delete all the stuff not needed and then do a search and save of the plates. It isn't automated, but as long as it is in the log file on my machine, I am ok with that.
Unfortunately, the price point for plate recognizer will probably push me to the OpenALPR route since I have so many plates.
I think for folks that have under 2,500 plates, this is a great simple add-on to BI.
I am thinking of ways to cut down the plates - maybe set up profiles to not pick up neighbors leaving and returning from work and then getting the night time. I always have the backup pics not sent to plate recognizer if needed.
Plate recognizer should look at the market and provide a comparable price point for the residential that would make this a no brainer. OpenALPR is $5/month and Plate recognizer is $50/month if you need more than 2,500 plates.
The integration with BI is so simple that at a lower price point, they would probably get more people. But if BI builds on the work folks have done here with OpenALPR, they will then take the market share. But maybe we are not the market for plate recognizer either?
Wow a lot more cars come past than I thought. I am already close to the halfway point of the 2,500 lookups
The simplicity of it is easy. And it is doing a fairly accurate job, even on plates that are total crap, but I know what they are. I was surprised and it gets them fast too.
It is simple enough to take the BI log and import into excel and quickly delete all the stuff not needed and then do a search and save of the plates. It isn't automated, but as long as it is in the log file on my machine, I am ok with that.
Unfortunately, the price point for plate recognizer will probably push me to the OpenALPR route since I have so many plates.
I think for folks that have under 2,500 plates, this is a great simple add-on to BI.
I am thinking of ways to cut down the plates - maybe set up profiles to not pick up neighbors leaving and returning from work and then getting the night time. I always have the backup pics not sent to plate recognizer if needed.
Plate recognizer should look at the market and provide a comparable price point for the residential that would make this a no brainer. OpenALPR is $5/month and Plate recognizer is $50/month if you need more than 2,500 plates.
The integration with BI is so simple that at a lower price point, they would probably get more people. But if BI builds on the work folks have done here with OpenALPR, they will then take the market share. But maybe we are not the market for plate recognizer either?
Thanks for the follow up.
I find their pricing strange. OK, let's say that they picked 2500/month as what they thought a hobbyist/home owner would get past their house. So let them get it for free. But that is only 83 'lookups'/day, which they say is any image sent even if there is no plate. I could see a few lookups being sent that are not even something with a plate on it. So maybe you get 75-80 real plates per day. In a rural setting or on a dead end street that would probably be fine. But in the present day urban or even suburban neighborhoods that may be too little. I have been trying to figure out how many would I need in a day. I am on a corner of a T-intersection. My street is on two different routes out of the subdivision so morning and afternoons there are a lot of cars coming and going. Then there are the delivery trucks, UPS, Fedex, Amazon, USPS. They are always zipping back an forth. Then the lawn guys. During school there are 6 busses a day. I have two cameras so depending on the travel path I get the same vehicle from both cameras. I don't think 80/day will do it for me.
But then they jump to 50k/month for $50, which is out of any home owner or hobbyist's price range, including mine. That is basically $1/thousand. So if they offered something like 10k/month for $10, or 15k/month for $15, I would go for that.
@wittaj How long did you spend setting it up? Is it difficult to do? Do you run it on the same machine as BI? Are you using SDK or Cloud? What version of BI are you running?
Well, it should have been under 5 minutes to set it up LOL, but for some reason it wasn't sending pics at first, so I kept trying different things like re-keying in the key code, rebooting machine, etc. I am guessing there must have just been a delay from registering until it would accept.
But the actual process for plate recognizer is easy - set up an account (just email and password) and it sends you an email, you open that and then get the code that you then copy/past into BI.
I am on BI 5.2.9.6.
It is on same machine as BI.
From what I can tell, you can only do cloud for the free version. Need to buy a separate license for SDK on-site.
From their website, you can log in on any computer and pull up the last 7 days of plates and images.
I think if those of us using and liking Plate Recognizer send them emails about their pricing, maybe they will adjust? Probably not, but it may be worth a try.
They already have the infrastructure, so getting $5 or $10/month from the hobbyist type instead of nothing has to help.
At that price point, every 5-10 new paying customers would equal 1 new customer at $50/month and word of mouth, etc. would probably make it worth it?
But again, maybe we are not their target client, and they feel like under 2,500 plates/month free is a good tag to hook people in. I mean that sounds like a lot of plates, but I am in a 35 home no outlet subdivision and am flying through that LOL.
If you haven't tried it, I would say go for it. Set-it up and then walk away for an hour. Should take you less than 5 minutes. If you don't like or don't want to use, you simply remove the code from BI.
But the actual process for plate recognizer is easy - set up an account (just email and password) and it sends you an email, you open that and then get the code that you then copy/past into BI.
I am on BI 5.2.9.6.
It is on same machine as BI.
From what I can tell, you can only do cloud for the free version. Need to buy a separate license for SDK on-site.
From their website, you can log in on any computer and pull up the last 7 days of plates and images.
I think if those of us using and liking Plate Recognizer send them emails about their pricing, maybe they will adjust? Probably not, but it may be worth a try.
They already have the infrastructure, so getting $5 or $10/month from the hobbyist type instead of nothing has to help.
At that price point, every 5-10 new paying customers would equal 1 new customer at $50/month and word of mouth, etc. would probably make it worth it?
But again, maybe we are not their target client, and they feel like under 2,500 plates/month free is a good tag to hook people in. I mean that sounds like a lot of plates, but I am in a 35 home no outlet subdivision and am flying through that LOL.
If you haven't tried it, I would say go for it. Set-it up and then walk away for an hour. Should take you less than 5 minutes. If you don't like or don't want to use, you simply remove the code from BI.
@wittaj Thanks for the full and quick reply! After last night's post I got to thinking. Just how many plates go by my house, really? So this morning I set my overview cam of the corner to grab a snapshot of each vehicle and put it in a separate folder. From 9:30am until noon I have had 63 real hits and 4 golf carts. I missed the folks going to work this morning. Figure I will run it for a couple of days and see how many I get.
I don't have my LPR cams set up yet. Still have to run the wires and install the camera box in the garden. It has proven to be a chore to get the cables from the attic over my garage into the beams that support the porch roof. I will not be able to get to that for about two weeks.
Thanks for the info.
I don't have my LPR cams set up yet. Still have to run the wires and install the camera box in the garden. It has proven to be a chore to get the cables from the attic over my garage into the beams that support the porch roof. I will not be able to get to that for about two weeks.
Thanks for the info.
Yes it is. I often sit out on the porch and play on my iPad and watch the world pass by. Some times it has not much traffic, then some times it's like a freeway.
Attachments
StratRider
Getting comfortable
the quick math does come to 80 plates or images per day in a 31 day month to stay under 2500 but then if you live in a state not requiring a front plate you need 2 cameras so cut that down to 40 plates/images per day that it will send - that's not much.
I have a cheap cam on each side of my house that I have also set crossing zones on - so when a car is traveling east - you could have it only get a snap or vid from a certain camera - same for the west side.
although that would let you do 80 per day - it seems that in my neighborhood (which I thought was pretty low traffic) - in reality I easily get more than that.
Currently I am doing it this way because I only have a 12x PTZ out front that is capable of catching plates and it is working fine this way. Although a lot more difficult using zones at night.
No doubt though that using 2 fixed 12x cams are in my future as I find minor problems with the PTZ method.
A car stopping in front of my house would never get plates logged as the PTZ would be looking down the road a bit waiting for the plate and would not be activated on a car leaving my house either.
Regardless, I may be able to justify the $5/camera/mo. for OpenALPR but I keep watching any LPR threads and hoping for a method directly logging plates to your own .csv file without ever sending them to another company.
It just seems that since OCR has been around a VERY long time - there should be a way to apply it on your own machine. OCR is done in a lot of areas but their costs are quite high also. Oh well.
I have a cheap cam on each side of my house that I have also set crossing zones on - so when a car is traveling east - you could have it only get a snap or vid from a certain camera - same for the west side.
although that would let you do 80 per day - it seems that in my neighborhood (which I thought was pretty low traffic) - in reality I easily get more than that.
Currently I am doing it this way because I only have a 12x PTZ out front that is capable of catching plates and it is working fine this way. Although a lot more difficult using zones at night.
No doubt though that using 2 fixed 12x cams are in my future as I find minor problems with the PTZ method.
A car stopping in front of my house would never get plates logged as the PTZ would be looking down the road a bit waiting for the plate and would not be activated on a car leaving my house either.
Regardless, I may be able to justify the $5/camera/mo. for OpenALPR but I keep watching any LPR threads and hoping for a method directly logging plates to your own .csv file without ever sending them to another company.
It just seems that since OCR has been around a VERY long time - there should be a way to apply it on your own machine. OCR is done in a lot of areas but their costs are quite high also. Oh well.
So this thread got me to thinking. I rally like Plate Recognizer since it seems so easy to set up and I have sent test captures to them and they all have been spot on.
So Wednesday I started having my corner overview cam count cars. I now have two full days of counts, Thursday and Friday. It is real interesting to see the hourly distribution. Nothing both days between 1 and 4 am. Numbers were fairly consistent by the hour between the days.
So I got 366 on Thursday and 382 on Friday. WAY MORE than I ever expected! These are just any vehicle that would have a plate. I did not count the several golf carts that go by. A lot of these cars will be logged by both the LPR cams once they are installed. Only the ones that enter the T from the East or leave the T toward the East would not get counted twice. While I don't have a good number on that, if I assume that half of them would only get counted once that gives me about 570 per day plus random triggers. So say 600 per day for a monthly total of about 18,000.
That is way above the 2500 free and way below the 50,000 for $50 per month. I wish they offered something like 25k/$25 or even 20k/$25. So I guess I may have to go the OPENALPR route at 2 cams for $10. It seems that is kind of difficult to set up and get reports/data back to your PC for analysis.
I still have time to think about this as I can not get to install the LPR cams for a few weeks. Bummer it will have to be done in the hottest part of the summer here.
So Wednesday I started having my corner overview cam count cars. I now have two full days of counts, Thursday and Friday. It is real interesting to see the hourly distribution. Nothing both days between 1 and 4 am. Numbers were fairly consistent by the hour between the days.
So I got 366 on Thursday and 382 on Friday. WAY MORE than I ever expected! These are just any vehicle that would have a plate. I did not count the several golf carts that go by. A lot of these cars will be logged by both the LPR cams once they are installed. Only the ones that enter the T from the East or leave the T toward the East would not get counted twice. While I don't have a good number on that, if I assume that half of them would only get counted once that gives me about 570 per day plus random triggers. So say 600 per day for a monthly total of about 18,000.
That is way above the 2500 free and way below the 50,000 for $50 per month. I wish they offered something like 25k/$25 or even 20k/$25. So I guess I may have to go the OPENALPR route at 2 cams for $10. It seems that is kind of difficult to set up and get reports/data back to your PC for analysis.
I still have time to think about this as I can not get to install the LPR cams for a few weeks. Bummer it will have to be done in the hottest part of the summer here.
Yea, I love the integration of Plate Recognizer and the accuracy has been amazing.
It is awesome that it puts the plates number on the alerts trigger image text information.
It is even better that it adds it to the log file so that you can simply import it into an excel file.
It is nice that they don't charge per camera, but not many are in a situation where 2,500 plates is enough. Maybe the end of no outlet street.
I am going to email them next week and see if they have considered a residential price that is more in line with OpenALPR.
If not, I will probably go thru the steps others have outlined in how to automate the OpenALPR.
The ability to log the plates into a database is what I am most interested in, and it seems like something always changes with OpenALPR requiring people to go in and make changes to keep that file going. That is where Plate Recognizer wins.
It is awesome that it puts the plates number on the alerts trigger image text information.
It is even better that it adds it to the log file so that you can simply import it into an excel file.
It is nice that they don't charge per camera, but not many are in a situation where 2,500 plates is enough. Maybe the end of no outlet street.
I am going to email them next week and see if they have considered a residential price that is more in line with OpenALPR.
If not, I will probably go thru the steps others have outlined in how to automate the OpenALPR.
The ability to log the plates into a database is what I am most interested in, and it seems like something always changes with OpenALPR requiring people to go in and make changes to keep that file going. That is where Plate Recognizer wins.
Have you settled upon a database? It would be nice to be able to do queries like 'show all plates for date/time range' or 'have you seen this plate?'. But at 18,000/month what could you use? That's a lot of lines in Excel. In a year you almost max out the number of lines you are allowed in Excel.
I don't know. Maybe I'll give PR a try and see how it goes. See if I feel I am getting worth out of $50/month.
I haven't got that far yet LOL.
But yea, I guess a database of some sort would probably be better in the long run.
I was testing this with obviously a small sample. But with it being basically a txt file, you could import it into anything.
$50/month is too much. At that point, if the Hikvision LPR camera zoom was enough for you, that camera would pay for itself quickly as it has LPR capability built in. Now how often you would have to download the data and that type of stuff I do not know as that was out of my price point when I first got into LPR and the zoom wasn't large enough.
Now being over a year with LPR, I wanted to go to the next level and get this type of database. Maybe in the next month or so I will try to steps outlined in the OpenALPR thread to auto download the data file and see how that goes.
But yea, I guess a database of some sort would probably be better in the long run.
I was testing this with obviously a small sample. But with it being basically a txt file, you could import it into anything.
$50/month is too much. At that point, if the Hikvision LPR camera zoom was enough for you, that camera would pay for itself quickly as it has LPR capability built in. Now how often you would have to download the data and that type of stuff I do not know as that was out of my price point when I first got into LPR and the zoom wasn't large enough.
Now being over a year with LPR, I wanted to go to the next level and get this type of database. Maybe in the next month or so I will try to steps outlined in the OpenALPR thread to auto download the data file and see how that goes.