OCR

OCR (Optical Character Recognition) with Google's AI technology (Cloud Vision API).

The Vision API can detect and extract text from images.

Install

1. Download Python 3.6+ or follow this guide from Google Cloud.

2. Clone or download this repository.

3. Install the required dependencies by running the following command in your terminal:

   !pip install google-cloud-vision google-auth.

Usage

usage: OCR [-h] [--url] [--document] [--languages LANGUAGES] [--full] [--confidence CONFIDENCE] [--key KEY] path

positional arguments:
  path                  path to image

optional arguments:
  -h, --help            show this help message and exit
  --url                 specify the path for an external image located on the Web (http:// or https://) or in Google Cloud Storage (gs://)
  --document            optimized for dense images
  --languages LANGUAGES, --language LANGUAGES
                        specify language hints from https://cloud.google.com/vision/docs/languages (comma separated)
  --full, --verbose     show full description (paragraphs, per-word confidence, boundaries...)
  --confidence CONFIDENCE
                        display possible mistakes for symbols with low confidence. Default: 0.6
  --key KEY             explicitly define the path to your service account JSON credentials

Authentication

Follow these instructions to set up a project with the Cloud Vision API enabled:

1. Select or create a Google Cloud Platform project. Project name suggestion: OCR

2. Enable Cloud Vision API for your project.

3. Create a service account and get your JSON credentials. Service account name suggestion: OCR

4. Make sure that billing is enabled for your project.

   Pricing is based on Google Cloud Vision API quota: 1,000 requests/month free

To authenticate your project you need to reference the service account JSON credentials you just downloaded. You have different options to do it, choose what you prefer:

service_account.json file

The JSON you downloaded in step 3 place it inside this repository folder.

--key parameter

Another option is to explicitly specify the --key parameter on every script execution:

OCR image.jpg --key "/path/to/service_account.json"

Environment variable

You can also set the GOOGLE_APPLICATION_CREDENTIALS environmental variable:

bash

Add to your .bash_profile file:

export GOOGLE_APPLICATION_CREDENTIALS="/path/to/service_account.json"

fish

Add to your config.fish file:

set -gx GOOGLE_APPLICATION_CREDENTIALS "/path/to/service_account.json"

Example

from File: TAJ.jpg

Result: output.txt