plate recognition for persian plates

In this project we extract plate text from persian plates.

Plateform

Windows (These scripts modified for windows version 10)

python >3.6

Directories

1- src --> main scripts folder

2- src/db --> db folder

3- data --> additional data

Installation

1- Create an virtual enviorment (You can use anaconda or virtualenv) and active the enviornment.

2- cd TO-PROJECT-FILES-FOLDER;

3- pip install -r reqierments.txt

(If you are living in the IRAN, for installation maybe you need to use a VPN.)

4- Head over to https://github.com/UB-Mannheim/tesseract/wiki and get the 32-bit or 64-bit version depending on your system architecture and install it like as other programs. (Or you can use direct path in your codes), you can use installation file in ~/data/ folder.

5- Put fas-tune-float.traineddata (in the ~/data/ folder) file in the C:\Program Files\Tesseract-OCR\tessdata path

6- Install the SQL Server drivers for your platform on the client system.

Troubleshooting

1- pyodbc.InterfaceError: This problem occured when you do not install odbc-driver or set not true drive. You can use the following cods for define your drivers and set true driver.

msa_drivers = [x for x in pyodbc.drivers() if 'ACCESS' in x.upper()]
print(msa_driver)

In the following an example of output is shown:

MS-Access Drivers : ['Microsoft Access Driver (*.mdb)', 'Microsoft Access Text Driver (*.txt, *.csv)']

and than change the following line in db.py:

driver = "Driver={Microsoft Access Driver (*.mdb, *.accdb)};"

to

driver = "Driver={Microsoft Access Driver (*.mdb)};

Usage

cd src
config parameters.py 
python main.py

Work stages

• Image capture by camera
• Plate recognition
  1. Read image and convert to gray scale
  2. Running threshold
  3. Running two types of erotion
  4. Extract contours by min and max area for extraction good contours
• OCR
  1. Running OCR over any contours
  2. Check the output is a plate numbers or not
  3. You can set minimum_persian_numbers_count and maximum_persian_numbers_count variables in the parameters.py.
• Saveing data in the database and save crop plates in the plate_images folder

Important notes

1- All logs saved in the logs folder. (logs/main.log)

2- When main.py run, temp and plate_images and logs folder will be created.

Database usage (seperatly)

I have created a database by the name "Plate_table" with three columns (image_path, plate_text, capture_date) you can change it with the following scripts by yourself.

from db import DB
db = DB()
cursor, conn = db.connect()

List of tabls

db.list_of_tables(cursor, conn)

Create a table

query = "CREATE TABLE Plate_table(image_path varchar(70), plate_text varchar(70), capture_date date)"
db.create_table(cursor, conn, query)

Remove a table

query = 'DROP TABLE Plate_table'
db.remove_table(cursor, conn, query)

Insert to a table

query = "INSERT INTO Plate_table ([image_path], [plate_text],[capture_date]) VALUES (?,?, ?)"
values = ("val", "text plate", datetime.datetime.now())
db.insert(cursor, conn, query, values)

Update a table

query = "UPDATE Plate_table SET plate_text = ? WHERE capture_date = ?"
values = ("update value", datetime.datetime(2021, 6, 11, 19, 52, 37))
db.update(cursor, conn, query, values)

Read rows of a table

query = 'select * from {}'.format("Plate_table")
print(db.select_rows(cursor, conn, query))