[Documentation] [TitleIndex] [WordIndex


The vision_opencv stack provides packaging of the popular OpenCV library for ROS. For information about the OpenCV library, please see the OpenCV main page at http://opencv.org/ links to complete documentation for OpenCV, as well as other OpenCV resources (like the bug tracker on http://code.opencv.org/)

For OpenCV vision_opencv provides several packages:

In order to use ROS with OpenCV, please see the cv_bridge package.

As of electric, OpenCV is a system dependency.

Using OpenCV in your ROS code

OpenCV2 is the official version supported on Indigo and Jade. To use it, you just need to add a dependency on opencv2 and find_package it in your CMakeLists.txt as you would for any third party package:

   target_link_libraries(my_awesome_library ${OpenCV_LIBRARIES})

You can also use OpenCV3: in that case, add a dependency to opencv3. But make sure that none of your dependencies depends on OpenCV2 (as you would get linked to both the OpenCVs which would most likely create a symbol conflict).

If you have OpenCV2 installed and ROS OpenCV3, OpenCV3 will be find_package-ed first. If you do not want to compile against OpenCV3 but still wish to have it installed, just find_package OpenCV2 as follows:

   find_package(OpenCV 2 REQUIRED)

Report an OpenCV specific Bug

If your issue is related to the OpenCV packaged in ROS (it is too old, you would like to see a backport in there ...), please file a bug for vision_opencv using the link provided at the top of this page.

For issues specific to OpenCV:




Since Indigo, there is a package for OpenCV3. Information about it is detailed at opencv3.

2024-07-20 14:46