Text extraction with GCP Cloud Vision API (Python3.6)

Introduction

I tried extracting text from images with GCP Cloud Vision API

Development environment

• Windows 10
• Anaconda
• Python 3.6
• OpenCV 4.4.0

Introduction

Refer to Detect text in images (https://cloud.google.com/vision/docs/ocr).

1. 1. Create a project from the Cloud Console.
2. Make sure billing is enabled.
3. 3. Enable the Vision API. Four. Set the authentication and the JSON file will be downloaded to your PC. Five. Set the environment variable GOOGLE_APPLICATION_CREDENTIALS to the JSON file path.
4. Open the anaconda prompt and create a Python 3.6 environment.

$ conda create -n py36 python=3.6
$ conda activate py36

7. 7. Install the library

$ pip install numpy
$ pip install pillow
$ pip install opencv-python
$ pip install --upgrade google-cloud-vision

8. 8. Let's run the code below

from google.cloud import vision
import io
import os
import cv2
import numpy as np
from PIL import ImageFont, ImageDraw, Image

def detect_text(image):
    """Detects text in the file."""
    client = vision.ImageAnnotatorClient()

    content = cv2.imencode(".png ", image)[1].tostring()
    tmp = vision.Image(content=content)
    response = client.text_detection(image=tmp)
    texts = response.text_annotations
    if response.error.message:
        raise Exception(
            '{}\nFor more info on error messages, check: '
            'https://cloud.google.com/apis/design/errors'.format(
                response.error.message))
    return texts

filename = "338px-Atomist_quote_from_Democritus.png "
root, ext = os.path.splitext(filename)
image = cv2.imread(filename, cv2.IMREAD_COLOR)
texts = detect_text(image)
fontpath ='C:\Windows\Fonts\meiryo.ttc'
font = ImageFont.truetype(fontpath, 10)
image_pil = Image.fromarray(image)
for text in texts:
    print(text.description)
    vertices = [(vertex.x, vertex.y) for vertex in text.bounding_poly.vertices]
    # cv2.putText(image, text.description, vertices[0], cv2.FONT_HERSHEY_PLAIN, 1, (255, 255, 255), 1, cv2.LINE_AA)
    # cv2.rectangle(image, vertices[0], vertices[2], (0, 255, 0))
    draw = ImageDraw.Draw(image_pil)
    w, h = draw.textsize(text.description, font = font)
    draw.text((vertices[0][0], vertices[0][1]-h), text.description, font=font, fill=(255, 255, 255, 0))
    # draw.text(vertices[0], text.description, font=font, fill=(255, 255, 255, 0))
    draw.rectangle((vertices[0], vertices[2]), outline=(0, 255, 0))
image = np.array(image_pil)
cv2.imshow("image", image)
cv2.imwrite(root+"_ocr"+ext, image)
cv2.waitKey(0)

input	PIL	PIL(OpenCV style)	OpenCV

With OpenCV, double-byte English (Japanese) is garbled, so I used PIL to display the text. Thank you for your hard work.