[Python] Use OCR to identify numbers of a water meter [closed]

I have been trying to solve this problem for quite some time, I have photos of the numbers of a water meter and I would like to obtain the numbers as text using Python and OCR or OpenCV, I tried with easyocr and pytesseract but without luck, maybe I have not clean enough the image before using the OCR software.

This is my image before modifying it:



After "cleaning it":



This is the code I'm using:

import easyocr
import os

AnaliticDir = 'Path_To_File'

reader = easyocr.Reader(['en'], gpu = True)

DownloadedFiles = os.listdir(AnaliticDir)
for i in range(len(DownloadedFiles)):
print(DownloadedFiles)
SRC_Image = AnaliticDir + "/" + DownloadedFiles
result = reader.readtext(SRC_Image, allowlist='0123456789',paragraph= True,decoder= 'beamsearch',batch_size=4,text_threshold=0.9)
for detection in result:
print(detection[1])


The resulting text after processing should be 0168390.

How can I improve or change my code so the OCR works well? Maybe I can do OpenCV object recognition with every character or something like that?

Continue reading...