OpenCV Text Area Detect

Searching text in a image, an opencv based text area detect tool.

1. Run on local

1.1 Verified environment

• Python 3.6
• OpenCV 3.2
• Tesseract 4.0 Alpha

1.2 Preparation

$ pip3 install opencv-python
$ pip3 install opencv-contrib-python

1.3 Detect text area only

$ python3 text_detect_wrap.py /path/to/image.jpg [out-put-folder]
$ open result.jpg   # or out-put-folder/result.jpg

1.4 Detect text content

$ bash text-detect.sh /path/to/image.jpg [out-put-folder]
$ cat /path/to/image/result.trim.txt   # or out-put-folder/result.trim.txt

2. Build and run the service via docker

2.1 Generate tesseract packages

$ docker build -t t4-pkg -f docker/Dockerfile.pkg docker
$ docker run --rm -it -v $(pwd)/docker:/pkg t4-pkg

2.2 Build base image

$ docker build -t t4-base -f docker/Dockerfile.base docker

2.3 Build service image

$ docker build -t t4 .

2.4 Start service

$ docker run -dt --name t4 -p 5000:5000 t4

2.5 View detect result in browser

URL: http://<IP>:5000/detect/<url-of-an-image-to-detect>