Python 10pines recibo OCR

detect.py

import cv2
import os
import sys
import numpy as np

def findTemplateInImage(img, templatePath):
    template = cv2.imread(templatePath, 0)

    # Apply template Matching
    res = cv2.matchTemplate(img, template, cv2.TM_SQDIFF_NORMED)
    min_val, max_val, min_loc, max_loc = cv2.minMaxLoc(res)

    # If the method is TM_SQDIFF or TM_SQDIFF_NORMED, take minimum
    return min_loc

img = cv2.imread(sys.argv[1],0)
subtotal = findTemplateInImage(img, os.environ['GAN_RECOG_WORKDIR'] + '/subtotal.template.png')
cuil = findTemplateInImage(img, os.environ['GAN_RECOG_WORKDIR'] + '/cuil.template.png')



# CUIL
##############################################################################
cuilImage = img[cuil[1]+15:cuil[1]+40, cuil[0]+150:cuil[0]+270]
cuilImageName = 'cuil.png'
cv2.imwrite(cuilImageName, cuilImage)
# cv2.imshow(cuilImageName, cuilImage)
# cv2.waitKey(0)
print('written cuil')

# SUBTOTAL
##############################################################################
subtotalImage = img[subtotal[1]+15:subtotal[1]+40, subtotal[0]+160:subtotal[0]+240]
subtotalImageName = 'subtotal.png'
cv2.imwrite(subtotalImageName, subtotalImage)
# cv2.imshow(subtotalImageName, subtotalImage)
# cv2.waitKey(0)
print('written subtotal')

# cv2.destroyAllWindows()

detect.sh

#/bin/bash
set -xeuo pipefail
export GAN_RECOG_WORKDIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"

function detect_single_recipe() {
  echo "Translating $1"

  TEMPDIR="$(mktemp -d --suffix=ganancias)"

  cp "$1" "$TEMPDIR"

  cd "$TEMPDIR"

  python "$GAN_RECOG_WORKDIR/detect.py" "$1"

  convert -units PixelsPerInch "cuil.png" -resample 300 "cuil.png"
  convert -units PixelsPerInch "subtotal.png" -resample 300 "subtotal.png"

  CUIL=$(tesseract "cuil.png" stdout | sed 's/[ \n]//g')
  SUBTOTAL=$(tesseract "subtotal.png" stdout | sed 's/[ \n]//g')

  cd -
  rm -rf "$TEMPDIR"

  echo "$SUBTOTAL" > "$CUIL.txt"
}

PDF_NAME=$(realpath "${1}")
WORKING_DIRECTORY=${1%.*}

mkdir -p "$WORKING_DIRECTORY"
cd "$WORKING_DIRECTORY"

pdftk "$PDF_NAME" burst

parallel --eta -j 4 'convert -density 150 {} -quality 90 {.}.png' ::: pg*.pdf
export -f detect_single_recipe
parallel --eta -j 4 'detect_single_recipe {}' ::: pg*.png