Correct, I would say this is the better way to do it. But if you want to be a purist and run the inference on the edge - I did that here. th-cam.com/video/qRFR0_FvkSs/w-d-xo.html
@@DataSlayerMedia Sweet projects. Thanks for sharing! I feel like we've moved from an era of "there's an app for everything" to the era of "there's an API for everything"
maybe it can be done on device with the right software or something just a little more powerful that a raspberry pi. Otherwise if the cloud service goes down or charges fees not the best.
How many images do you send to the plate recognition system send per second? I mean if you split 24 frames per second do you send 24 images to the API? Thank you in advance.
Hey man, great video! I'm trying to recreate this for a school project but i need help with 2 things. 1. When i run the code it just says "waiting for change in flags.txt", what do I need to do? 2. Where do I put the API key from the website? Thanks!
Did you test the maximum offset angle for your LPR? And I assume this will not work at night without some version of IR illumination and a night vision camera?
Is there any to set this up to run with a Canon DSLR or a way to get high res or pro grade photos? Or take the photos with the Canon, and using syncthings to auto upload the photos from my phone to my PC or a Raspberry Pi and have them ran by the recognitioning and put in a database? I take a lot of car photos and I would like to have a database that inputs the tag number, and vehicle make and model either from the photo or like a CarFax type service that I can sort a few different ways.
Hi, the subtitles are more annoying than helpful. This is not a format my brain can process easily. It would be nice if you just create youtube subtitles or let youtube create them, it would be much more legible that way! thanks! congrats with the project!
Sorry but I strongly disagree. Sincerely, a Deaf guy who can't hear. hard subs > automatic CC Besides, it is up to the content creator - as long I can enable CC on TH-cam I am satisfied, but I am always happier when I see hard subs.
I tend to agree with you... BUT if I look at my retention graph data it's very clear that the videos with burned in subtitles retain viewership over the vids without it. I will continue to experiment though and appreciate the feedback.
@@DataSlayerMedia hi. The main criticism is the fact that those burned in subs just go way to fast. Coming from a country where almost everything is subbed, my brain parses them in two lines, not in 1 word.
@@BertClaes I re-read your comment and I think I agree with you - halfway. I always prefer hard subs, BUT you do have a point, I’d like to see longer/comprehensible sentences than short flashing words. But I never get bothered because deaf people are trained since birth to read at laser speed because we read and read captions all day every day.
Could this be adapted to read speed limit signs (from within the car) as the car passes them? I don't think OCR is needed you could just match images of different signs at 30, 40 etc mph.
I need this for detecting license plates entering snd exiting a condo complex mounted on a pole. Additionally need to have solar and only take pictures on motion any suggestions what parts are needed?
As you do nothing on the Pi itself, I would say you owe somebody 100.00 EUR. Because this is like saying that you can make a blind person say the color of a car passing by and all you do is tell him the color of the car. And perhaps instead of send the person 100 USD, just send her 3 of these 35 USD Pi systems and ask 5 USD back.
Want to run the ML inference on the edge? 👉th-cam.com/video/qRFR0_FvkSs/w-d-xo.html
So the RPi is just capturing the images and sending them to the cloud, not actually doing the recognition processing...
Correct, I would say this is the better way to do it. But if you want to be a purist and run the inference on the edge - I did that here. th-cam.com/video/qRFR0_FvkSs/w-d-xo.html
@@DataSlayerMedia Sweet projects. Thanks for sharing! I feel like we've moved from an era of "there's an app for everything" to the era of "there's an API for everything"
Most of the work is done by the cloud service not the pie. The cloud service has massive computation which is needed
maybe it can be done on device with the right software or something just a little more powerful that a raspberry pi. Otherwise if the cloud service goes down or charges fees not the best.
How many images do you send to the plate recognition system send per second? I mean if you split 24 frames per second do you send 24 images to the API? Thank you in advance.
Phooey! Cloud plate recognition service? Then it is NOT an RPi plate scanner.
You weren't arrested but surely scared some drivers
Hey man, great video! I'm trying to recreate this for a school project but i need help with 2 things. 1. When i run the code it just says "waiting for change in flags.txt", what do I need to do? 2. Where do I put the API key from the website? Thanks!
Hi, I have the same problem. Did you get to solve it? I'm also stuck in waiting for change in flag.txt
@@aldofelix2002 did you guys resolve this issue?
how close the camera has to be to capture a clear plate number? I setup camera module 3 in my drive way but it doesn't see plate clearly.
Did you test the maximum offset angle for your LPR? And I assume this will not work at night without some version of IR illumination and a night vision camera?
Is there any to set this up to run with a Canon DSLR or a way to get high res or pro grade photos? Or take the photos with the Canon, and using syncthings to auto upload the photos from my phone to my PC or a Raspberry Pi and have them ran by the recognitioning and put in a database? I take a lot of car photos and I would like to have a database that inputs the tag number, and vehicle make and model either from the photo or like a CarFax type service that I can sort a few different ways.
Hi, the subtitles are more annoying than helpful. This is not a format my brain can process easily. It would be nice if you just create youtube subtitles or let youtube create them, it would be much more legible that way! thanks! congrats with the project!
Sorry but I strongly disagree.
Sincerely, a Deaf guy who can't hear.
hard subs > automatic CC
Besides, it is up to the content creator - as long I can enable CC on TH-cam I am satisfied, but I am always happier when I see hard subs.
I tend to agree with you... BUT if I look at my retention graph data it's very clear that the videos with burned in subtitles retain viewership over the vids without it. I will continue to experiment though and appreciate the feedback.
@@DataSlayerMedia hi. The main criticism is the fact that those burned in subs just go way to fast. Coming from a country where almost everything is subbed, my brain parses them in two lines, not in 1 word.
i agree, these subtitles are pretty annoying
@@BertClaes I re-read your comment and I think I agree with you - halfway. I always prefer hard subs, BUT you do have a point, I’d like to see longer/comprehensible sentences than short flashing words.
But I never get bothered because deaf people are trained since birth to read at laser speed because we read and read captions all day every day.
that "bro" destroyed me too!
Could this be adapted to read speed limit signs (from within the car) as the car passes them? I don't think OCR is needed you could just match images of different signs at 30, 40 etc mph.
I need this for detecting license plates entering snd exiting a condo complex mounted on a pole. Additionally need to have solar and only take pictures on motion any suggestions what parts are needed?
What's the case name?
www.amazon.com/gp/product/B07JBB9QSB/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1
As you do nothing on the Pi itself, I would say you owe somebody 100.00 EUR. Because this is like saying that you can make a blind person say the color of a car passing by and all you do is tell him the color of the car. And perhaps instead of send the person 100 USD, just send her 3 of these 35 USD Pi systems and ask 5 USD back.
Where can you buy any raspberry pi for $35?
You can buy it for $35 off their website, when they come back in stock... www.raspberrypi.com/products/raspberry-pi-4-model-b/
Microcenter
This is terrifying.
Like this video. I follow you.
This has to be the worst scripting and acting on TH-cam