- Merge pull request #191 from patrickelectric/kinetic cv2_to_imgmsg: step must be int
- cv2_to_imgmsg: step must be int Signed-off-by: Patrick José Pereira <[email protected]>
- Contributors: Patrick José Pereira, Vincent Rabaud
- fix endianness issues
- Contributors: Vincent Rabaud
- add version_gte for opencv3 @vrabaud If you'll update opencv3 version as discussed in https://discourse.ros.org/t/opencv-3-3/2674/4, I think we'd better to add 'version_gte' tag so that apt-get install ros-kinetic-cv-bridge also pulls openv3.3 from repository, to avoid API breaking issue between opencv2 and opencv3.
- Simplify the dependency components of cv_bridge Fixes #183
- Fixes #177 The Python bridge was wrong on OpenCV2 with mono8 (and any Mat with only two dimensions btw). Took the official Python bridge from OpenCV.
- Add missing test file This fixes #171
- Properly deal with alpha in image compression. That fixes #169
- Silence warnings about un-used variables
- export OpenCV variables
- Contributors: Kei Okada, Victor Lamoine, Vincent Rabaud
- properly find Boost Python 2 or 3 This fixes #158
- Contributors: Vincent Rabaud
- Use api in sensor_msgs to get byte_depth and num channels
- Implement cpp conversion of N channel image This is cpp version of ros-perception/vision_opencv#141, which is one for python.
- Fill black color to depth nan region
- address gcc6 build error in cv_bridge and tune With gcc6, compiling fails with stdlib.h: No such file or directory, as including '-isystem /usr/include' breaks with gcc6, cf., https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70129 This commit addresses this issue for cv_bridge in the same way it was done in the commit ead421b8 [1] for image_geometry. This issue was also addressed in various other ROS packages. A list of related commits and pull requests is at: ros/rosdistro#12783 [1] https://github.com/ros-perception/vision_opencv/commit/ead421b85eeb750cbf7988657015296ed6789bcf Signed-off-by: Lukas Bulwahn <[email protected]>
- cv_bridge: Add missing test_depend on numpy
- Contributors: Kentaro Wada, Lukas Bulwahn, Maarten de Vries
- Specify background label when colorizing label image
- Adjust to arbitrary image channels like 32FC40 Proper fix for #141
- Remove unexpectedly included print statement
- Contributors: Kentaro Wada, Vincent Rabaud
- split the conversion tests out of enumerants
- support is_bigendian in Python Fixes #114 Also fixes mono16 test
- Support compressed Images messages in python for indigo - Add cv2_to_comprssed_imgmsg: Convert from cv2 image to compressed image ros msg. - Add comprssed_imgmsg_to_cv2: Convert the compress message to a new image. - Add compressed image tests. - Add time to msgs (compressed and regular). add enumerants test for compressed image. merge the compressed tests with the regular ones. better comment explanation. I will squash this commit. Fix indentation fix typo mistage: from .imgmsg_to_compressed_cv2 to .compressed_imgmsg_to_cv2. remove cv2.CV_8UC1 remove rospy and time depndency. change from IMREAD_COLOR to IMREAD_ANYCOLOR. - make indentaion of 4. - remove space trailer. - remove space from empty lines. - another set of for loops, it will make things easier to track. In that new set, just have the number of channels in ([],1,3,4) (ignore two for jpg). from: ros-perception/vision_opencv#132 (comment) - keep the OpenCV error message. from: ros-perception/vision_opencv#132 (comment) add debug print for test. add case for 4 channels in test. remove 4 channels case from compressed test. add debug print for test. change typo of format. fix typo in format. change from dip to dib. change to IMREAD_ANYCOLOR as python code. (as it should). rename TIFF to tiff Sperate the tests one for regular images and one for compressed. update comment
- Add CvtColorForDisplayOptions with new colormap param
- fix doc jobs
- Add python binding for cv_bridge::cvtColorForDisplay
- Don't colorize float image as label image This is a bug and image whose encoding is other than 32SC1 should not be colorized. (currently, depth images with 32FC1 is also colorized.)
- Fix compilation of cv_bridge with opencv3 and python3.
- Contributors: Kentaro Wada, Maarten de Vries, Vincent Rabaud, talregev
- depend on OpenCV3 only
- Contributors: Vincent Rabaud
- Fix my typo
- Remove another eval Because cvtype2_to_dtype_with_channels('8UCimport os; os.system("rm -rf /")') should never have a chance of happening.
- Remove eval, and other fixes Also, extend from object, so as not to get a python 2.2-style class, and use the new-style raise statement
- Contributors: Eric Wieser
- clean up the doc files
- fix a few warnings in doc jobs
- Contributors: Vincent Rabaud
- fix OpenCV3 build
- Describe about converting label to bgr image in cvtColorForDisplay
- Convert label to BGR image to display
- Add test for rgb_colors.cpp
- Add rgb_colors util
- Update doc for converting to BGR in cvtColorForDisplay
- Convert to BGR from any encoding
- Refactor: sensor_msgs::image_encodings -> enc
- Contributors: Kentaro Wada, Vincent Rabaud
- deal with endianness
- add cvtColorForDisplay
- Improved efficiency by using toCvShare instead of toCvCopy.
- Add format enum for easy use and choose format.
- fix compilation warnings
- start to extend the cv_bridge with cvCompressedImage class, that will convert from cv::Mat opencv images to CompressedImage ros messages and vice versa
- Contributors: Carlos Costa, Vincent Rabaud, talregev
- Simplify some OpenCV3 distinction
- fix tests
- fix test under OpenCV3
- Remove Python for Android
- Contributors: Gary Servin, Vincent Rabaud
- check that the type is indeed a Numpy one This is in response to #51
- Contributors: Vincent Rabaud
- chnage the behavior when there is only one channel
- cleanup tests
- Contributors: Vincent Rabaud
- get code to work with OpenCV3 actually fixes #46 properly
- Contributors: Vincent Rabaud
- Fix #42
- Contributors: Libor Wagner
- Correct dependency from non-existent package to cv_bridge
- Contributors: Isaac Isao Saito
- Add depend on python for cv_bridge
- Contributors: Scott K Logan
- fixes #34
- Contributors: Vincent Rabaud
- remove deprecated API and fixes #33
- fix OpenCV dependencies
- Contributors: Vincent Rabaud
- fix python 3 error at configure time
- Contributors: Dirk Thomas
- update changelog
- Find NumPy include directory
- Contributors: Brian Jensen, Vincent Rabaud
- fix compilation on older NumPy
- Contributors: Vincent Rabaud
- bump changelog
- Fixed issue with image message step size
- fix crash for non char data
- fix #26
- Contributors: Brian Jensen, Vincent Rabaud
- fix bad image check and improve it too
- Contributors: Vincent Rabaud
- fixes #25
- Contributors: Vincent Rabaud
- fixes #20
- Contributors: Vincent Rabaud
- fixes #22
- fixes #17
- check for CATKIN_ENABLE_TESTING
- fixes #16
- update email address
- Contributors: Lukas Bulwahn, Vincent Rabaud
- make sure conversion are applied for depth differences
- Contributors: Vincent Rabaud
- fix installation of the boost package
- Contributors: Vincent Rabaud
- Link against PTYHON_LIBRARIES
- Contributors: William Woodall
- use CATKIN_DEVEL_PREFIX instead of obsolete CATKIN_BUILD_PREFIX
- Contributors: Dirk Thomas
- add licenses
- fixes #5 by removing the logic from Python and using wrapped C++ and adding a test for it
- fix a bug discovered when running the opencv_tests
- use some C++ logic
- add a Boost Python module to have the C++ logix used directly in Python
- Contributors: Vincent Rabaud
- add conversion from Bayer to gray
- Contributors: Vincent Rabaud
- use the reverted isColor behavior
- Contributors: Vincent Rabaud
- use the catkin macros for the setup.py
- fix #3
- Contributors: Vincent Rabaud
- buildtool_depend catkin fix
- CMakeLists.txt clean up.
- Contributors: William Woodall
- fix issue #1
- Cleanup of package.xml
- Contributors: Vincent Rabaud, William Woodall
- fix the bad include folder
- Contributors: Vincent Rabaud
- fix dependencies
- Contributors: Vincent Rabaud
- fix some dependencies
- add rosconsole as a dependency
- fix missing Python at install and fix some dependencies
- Contributors: Vincent Rabaud
- add missing stuff
- make sure we find catkin
- Contributors: Vincent Rabaud
- move the test to where it belongs
- fix the tests and the API to not handle conversion from CV_TYPE to Color type (does not make sense)
- comply to the new Catkin API
- backport the YUV422 bug fix from Fuerte
- apply patch from https://code.ros.org/trac/ros-pkg/ticket/5556
- Contributors: Vincent Rabaud
- remove dependencies to the opencv2 ROS package
- Contributors: Vincent Rabaud
- make sure the include folders are copied to the right place
- Contributors: Vincent Rabaud
- be more compliant to the latest catkin
- added catkin_project() to cv_bridge, image_geometry, and opencv_tests
- Contributors: Jonathan Binney, Vincent Rabaud
- remove things that were marked as ROS_DEPRECATED
- Contributors: Vincent Rabaud
- catkinized opencv_tests by Jon Binney
- catkinized cv_bridge package... others disable for now by Jon Binney
- remove the version check, let's trust OpenCV :)
- revert the removal of opencv2
- vision_opencv: Export OpenCV flags in manifests for image_geometry, cv_bridge.
- finally get rid of opencv2 as it is a system dependency now
- bump REQUIRED version of OpenCV to 2.3.2, which is what's in ros-fuerte-opencv
- switch rosdep name to opencv2, to refer to ros-fuerte-opencv2
- added missing header
- Added constructor to CvImage to make converting a cv::Mat to sensor_msgs::Image less verbose.
- cv_bridge: Added unit test for #5206
- cv_bridge: Applied patch from mdesnoyer to fix handling of non-continuous OpenCV images. #5206
- Adding opencv2 to all manifests, so that client packages may not break when using them.
- baking in opencv debs and attempting a pre-release
- cv_bridge: Support for new 16-bit encodings.
- cv_bridge: Deprecate old C++ cv_bridge API.
- cv_bridge: Correctly scale for MONO8 <-> MONO16 conversions.
- cv_bridge: Fixed issue where pointer version to toCvCopy would ignore the requested encoding (http://answers.ros.org/question/258/converting-kinect-rgb-image-to-opencv-gives-wrong).
- fixed doc build by taking a static snapshot
- cv_bridge: Marking doc reviewed.
- cv_bridge: Tweaks to make docs look better.
- cv_bridge: Added cvtColor(). License notices. Documented that CvBridge class is obsolete.
- cv_bridge: Added redesigned C++ cv_bridge.
- Doc cleanup
- Trigger doc rebuild
- mono16 -> bgr conversion tested and fixed in C
- Added Ubuntu platform tags to manifest
- Handle mono16 properly
- Raise exception when imgMsgToCv() gets an image encoding it does not recognise, #3489
- Remove use of deprecated rosbuild macros
- Fixed example
- cv_bridge split from opencv2
- Contributors: Vincent Rabaud, ethanrublee, gerkey, jamesb, mihelich, vrabaud, wheeler