Only released in EOL distros:
The text_locator package
- Maintainer: Vojtech Novak <vonovak AT gmail DOT com>
- Author: Vojtech Novak <vonovak AT gmail DOT com>
- License: MIT
- Source: git https://github.com/vonovak/text_locator.git (branch: master)
text_locator is a ROS package that can detect and recognize text in images from the MS kinect / Asus Xtion sensors. Also, it can give you spatial information about the text, ie. where the text is in the 3D space, with respect to the sensor.
example usage video:
Install the text_locator package from the github repo (see here). Then download the CCV library from here, and compile it. The result will be a static library file, libccv.a. Copy this file to the /lib folder of the package.
To compile CCV, extract the tarball, cd to [extractedCCV]/lib folder, run ./configure followed by make. When configuring, the 'swscale' is the only component that is needed. The other compilation options (enable SSE2, use openmp etc..) are optional.
roslaunch text_locator text_locator.launch
The output of the node is advertised on text_locator_topic.
More detailed documentation is included in the git repo's readme file here.