tcanvn / driver-license-ocr

Driver-License-OCR Dependncy

Tesseract, freetype, imagemagick@6, ghostscript, openCV, python3. Django

What I achieved

• Login(authenticate)

  • Only authenticated users are able to use this service

• Upload image(jpg, png, jpeg) or pdf file(wand)

  • Use wand convert pdf to image

• Preprocess the picture(openCV, pillow)

  • Border Removal(Select image area with driver information)
  • Brightness Adjustment(BGR2HSV)
  • Smooth the picture(GaussianBlur)
  • Color Change(BGR2Gray)
  • Threshold Change(THRESH_BINARY/black and white)
  • Noise remove(erode and dilate)

• Recognize the picture(pytesseract)

  • Tesseract extract words from the processed image

• Get name and License numebr(rx)

  • Regular expression pick up the driver name and license nnumber

• Create the picture/pdf(wand, pypdf2, reportlab.pdfgen, zipfile)

  • Write driver name and number with Helvetica-Bold 14 size at correct position as Watermark
  • Merge Watermark and background pdf(create new pdf file)
  • Pdf to Image by using wand(optional)
  • Compress output files with zipfile

• Modify pages

  • bootstrap and css