@Gymratz please email me at your earliest convenience. Some changes were made to the API in a good way and I just want to get your software up to date with the changes.
I was looking through the GitHub repository and noticed a few things.
Thanks,
Tyler
Awesome, I will definitely keep that in mind. Current project I have been working on is to create suggestions on possible plates for partial hits within the database I have created.
Awesome, I will definitely keep that in mind. Current project I have been working on is to create suggestions on possible plates for partial hits within the database I have created.
Using that paired with LPR would be a game changer especially paired with information from the database. I have had a small testing setup due to @Gymratz software pushing my interest into LPR and have had success results with live footage/images.
But this is only with New Jersey License Plates as that is where I currently reside. I will look more into Rekor and learn from they're development to start my task on creating my own LPR API and hopefully beat the price they offer.
Easily something we could all contribute and create, the only issue is do we go web based or do we go local?You can find on GitHub the OpenALPR open source material that was out there until Rekor bought the creator out.
That programming is like two years old, but certainly enough people are here that collectively we could figure something out between the AI advancements and all the plate data that folks here have to train models.
Obviously it would take a lot to compete with Rekor, but if someone developed an opensource grass roots system created with our own plate images, it would be good enough for most of us. I think most of us would be happy if we had a plate reader that could read our home state and the few surrounding states. Rekor obviously has the whole country and other countries that we indirectly pay for the development, but I have no need for Argentina plates LOL.
Easily something we could all contribute and create, the only issue is do we go web based or do we go local?
I suggest web based, because as each user trains a the AI within their residence it would train everyone within the US. Like a daisy chain..
I'm going to look more into what you mentioned about OpenALPR open source, and see what I can do.
I was originally using OCR in PHP to gather text from image and it wasn't too bad. I would have to dig deeper to clean up the images for night plates with either license plate lights, IR and or street lights to be able to be read without issue.
Yes, I remember that!I had sent you earlier how we deal with getting night plates - run fast shutter (1/2,000) and blast the reflective plate with IR. All we get is the head/taillight and plates, but that is all we need!
For most of us, our night images produce a cleaner plate than the daytime LOL.
Perfect! Thank you, I appreciate it!Here is one representative sample I post here when NOOB comes asking about camera for plates.
I am sure several here can DM you some representative samples from their cams.
View attachment 110723
Oh my god... that is awesome!!If you need more, I'd happy give you access to all my 200K plates I've collected
View attachment 110732