anpr with python+yolo+opencv
Budget: $250 – $750 USD
requirements:
- train the yolo model using my images
- deploy the solution for running in a mini pc with celeron n4500, 8gb ram and 240 gb ssd nvme
- deploy the solution running into a ubuntu 22 (not vm)
- Has a whitelist and blacklist file with ocr for car plates:
*the car plate into whiltelist and blacklist will be only lettes and numbers: e.g. TVZ763A, 15F993, etc
*whitelist file are the car plates who the anpr will open the gate or barrier
*blacklist file are the car plates who the anpr will not open the gate or barrier
*These 2 files are going to be updated 1 time per month
- use 2 ip cam to get the cars video for take plates ( cam1 for cars in and cam1 for cars out)
- Use a 2CH IP relay module to send the opening pulse to open corresponding car barrier/gate
- write into a public html daily file for all cars plates detected: /var/www/html/anpr (on local pc)
- date and time
- IN or OUT
- ocr of number car plate detected
- status ( OK = car plate into whitelist | NOK = car plate into blacklist | UNK = car plate Unknown )
- small car photo
* each daily file must be in its corresponding folder for the current month
* All files and images of each month, must remain 2 additional month for consultation,
then all files of that month must be deleted.
e.g., all January files must remain during February and march and must be deleted on April 1,
and so on every month
- the anpr solution must be running offline
- train the yolo model using my images
- deploy the solution for running in a mini pc with celeron n4500, 8gb ram and 240 gb ssd nvme
- deploy the solution running into a ubuntu 22 (not vm)
- Has a whitelist and blacklist file with ocr for car plates:
*the car plate into whiltelist and blacklist will be only lettes and numbers: e.g. TVZ763A, 15F993, etc
*whitelist file are the car plates who the anpr will open the gate or barrier
*blacklist file are the car plates who the anpr will not open the gate or barrier
*These 2 files are going to be updated 1 time per month
- use 2 ip cam to get the cars video for take plates ( cam1 for cars in and cam1 for cars out)
- Use a 2CH IP relay module to send the opening pulse to open corresponding car barrier/gate
- write into a public html daily file for all cars plates detected: /var/www/html/anpr (on local pc)
- date and time
- IN or OUT
- ocr of number car plate detected
- status ( OK = car plate into whitelist | NOK = car plate into blacklist | UNK = car plate Unknown )
- small car photo
* each daily file must be in its corresponding folder for the current month
* All files and images of each month, must remain 2 additional month for consultation,
then all files of that month must be deleted.
e.g., all January files must remain during February and march and must be deleted on April 1,
and so on every month
- the anpr solution must be running offline