- .travis.yml: add noetic test (#108)
- support oepncv4 for face detection launch/test files
- convert img_gray every loop, this fixes [ERROR] [1597757072.115502734]: Image processing error: Matrix operand is an empty matrix. checkOperandsExist ../modules/core/src/matrix_expressions.cpp 23 errors
- check CV_VERSION_VERSION > 4
- add g++ static to package.xml
- use python3 for noetic
- .travis.yml : add CHECK_PYTHON3_COMPILE
- .travis.yml: add noetic test
- fix travis (#106)
- add more package to CATKIN_DEPENDS
- add image_view to run_depends
- add hot fix for hydro test (#98)
- add doublequote arount $TEST
- add hot fix for hydro test
- Improved variable names and comments (#95)
- Improved variable names and comments: The comments and variable names were the opposite of what was functionally happening in the code. No functional change to this commit, just better readability and maintainability.
- setup EoL repository (#96)
- Contributors: Gus Crowards, Kei Okada
- support catkin_lint and clang-format tests in travis.yml (#93)
- override is not supported gcc4.6 (12.04), remove this fix and add NOLINT
- clang-tidy code need c++11
- fix code by run-clang-tidy -fix
- fix format by clang-format
- fix CMakeLists.txt and package.xml for catkin_lint
- support catkin_lint and clang-format tests
- add queue_size parameter to all nodes, see #83 (#92)
- add queue_size arg to launch files
- sometimes simple_example_test fails with 'average rate (36.121Hz) exceeded maximum (35.000Hz)'
- add queue_size parameter to all nodes, see #83
- add melodic badge
- Contributors: Furushchev, Hironori Fujimoto, Kei Okada, higashide, iory, moju zhao
- add melodic badge
- Add lk flow params trackbar (#78)
-
- Remove duplication of add_library for simple_flow (#88)
-
${_opencv_apps_nodelet_cppfiles} adds simple_flow to library but also $ {${PROJECT_NAME}_EXTRA_FILES} does same thing
- Do not pefrom face recognition process without the trained data (#91)
- Add warning logger to prompt face data tranining
- Add a check to decide whether to perform the face recognition in callback function, according to the content of th
- fback_flow: add option to set 'queue_size' (#83)
- travis.yml: add melodic and remove jade (#84)
- [face_detection.launch] Fixed path of haarcascade xml for OpenCV-3.3.1 (#79)
- Contributors: Yuki Furuta, Hironori Fujimoto, Kei Okada, Taichi Higashide, Iory Yanokura, Moju Zhao
- Fix namespace and pkg name of nodelets (Closes (#21)) (#74) Fix namespace and pkg name of nodelets
- Add pyramids_nodelet (#37) * use toCvCopy instead of CvShare in adding_images
- adding_images_nodelt :support different size of images (#57)
- fix contour moment program (#66) * contour_moments_nodelet.cpp: remove redundant codes, use input encoding * contour_moments_nodelet.cpp: sort contours by the area * contour_moments_nodelet.cpp: remove tailing NR from NODELET_INFO
- fix for opencv 3.3.1 (#71) * fix launch/test fiels for opencv3.3 * goodFeaturesTrack takes useHarriesDetector == false * opencv 3.3.1 has newer FaceRecognizer
- Contributors: Kei Okada, Iori Yanokura
- [src/node/standalone_nodelet_exec.cpp.in] workaround for freezing imshow on kinetic (#67) * use ros::param::set instead of ros::NodeHandle("~"), that did not output NODELET_INFO * workaround for freezing imshow on kinetic
- [launch/hough_circles.launch] Corrected a typo and applied the node_name argument (#69 )
- [face_recognition] add nodelet / script / message files for face recognition (new) #63 from furushchev/face-recognition-new
- add face_recognition nodelet / test cfg/FaceRecognition.cfg launch/face_recognition.launch scripts/face_recognition_trainer.py src/nodelet/face_recognition_nodelet.cpp
- [Face.msg] add label / confidence for face recognition
- [CMakeLists.txt] remove duplicate msg: RectArrayStamped.msg
- cfg/*.cfg : Set useless use_camera_info flag to false in default (#58 )
- Contributors: Kei Okada, Kentaro Wada, Yuki Furuta, wangl5
- New Nodes
- [color_filter_nodelet.cpp] Add color_filter nodelet (#48) * use BGR2HSB, support H from 0-360 and 350 - 360+a * Unified hsl -> hls * Add hsv_color_filter test * Modified hls_color_filter's test paramter. Extracting skin color.
- [corner_harris_nodelet.cpp] Add corner-harris (#38 )
- [discrete_fourier_transform_nodelet.cpp] Add discrete_fourier_transform_nodelet (#36 )
- New Feature
- [face_detection_nodelet.cpp] publish face roi image (#40 )
- [face_detection_nodelet.cpp] fix: use encoding BGR8 on conversion from cv::Mat to sensor_msgs/Image
- Fix / Improvement
- [adding_images_nodelet.cpp] Fix AddingImages (#52)
- CvtColorForDisplay is not supported until 1.11.9 (hydro)
- CvtColorForDisplayOptions is supported in 1.11.13
- Rename topic ~info to camera_info for consistency
- Do dynamic scaling for float/double images
- Support adding images whose encodings are same kind, For example adding rgb8 + bgr8
- display using cvtColorForDisplay
- Clarify with auto_beta for auto beta settings
- Check input encoding consistency
- Add arbitrary dtype images
- AddingImages: enable to set beta param if use_data is true
- [face_detection] add test for face_detection/face_image topic (#49) * test/CMakeLists.txt : skip face_detection.test * [test/test-face_detection.test] add test for face_image
- [.travis.sh] bugfix: test for opencv3 #45 * [.travis.sh] bugfix: use --upstream for rosinstall_generator to fetch not only metapackage - [.travis.sh] run test only opencv_apps package (not dep packages) - [.travis.sh] build compressed_image_transport from source if opencv3 is enabled - [package.xml] use compressed_image_transport for test_depend instead of meta package image_transport_plugins
- [doc] Better package description. (#43)
- watershed_segmentation_nodelet.cpp : Fix typo in warnnige message (#34)
- Create README.md
- Contributors: Isaac I.Y. Saito, Kei Okada, Kentaro Wada, Yuki Furuta, Iori Yanokura
- Force convert to bgr for display (#30)
- add include sensor_msgs/image_encodings for old image_encodings
- force conver to bgr8 using sensor_msgs::image_encodings::BGR8
- Add more nodes from opencv sample codes
- Add launch files for opencv_apps nodes
- separate launch and test files (#20)
- Add hydro travis testing (#22)
- test/CMakeLists.txt : catkin_download_test_data not working with DESTINATION . for hydro
- cv_bridge before 1.11.9 does not suport CompressedImage in cv_bridge
- lk_flow : need to explicitly include sensor_msgs/image_endcodings.h
- simple_compressed_example_nodelet.cpp : need to include sensor_msgs/CompressedImage explicitly on hydro
- .travis.yml : add hydro testing
- Minor Fixes
- update gitignore to avoid test png data (#28)
- fix hough_circles for input frame color (#13 )
- CMakeLists.txt update list of opencv tutorial codes (#25)
- fix face_detection.launch to accept args for cascade xml for opencv3 (#20)
- CMakeLists.txt : add install rule for launch (#20)
- add launch/.launch files (from test/.test) to reuse launch files (#20)
- CMakeLists.txt: on roslaunch 1.11.1, roslaunch_add_file check fails with unsupported doc attributes (#20)
- Add test for simple_example / simple_compressed_example (#24)
- add retry for simple_example/simple_compressed_example test, not sure why it fails.. on travis
- package.xml : add image_transport_plugins to test_depend for republish node in test-simple_compressed_example.test
- add test for simple_example/simple_compressed_example
- simple_example_nodlet.cpp / simple_compressed_example_nodelet.cpp : support debug_view param
- .travis.sh : add catkin_test_results --verbose
- Support kinetic on travis (#15)
- test/test-face-detection.test : add haarcascade data from opencv3 package directory
- use docekr to run trusty/xenial .travis.sh
- Modified enabling use_camera_info by rosparam (#18)
- Enabling dynamic_reconfigure in private nodelet handler
- Enable to set min_distance_between_circles param, publish debug message (#14)
- hough_circles : fix to set dp_int to dp
- hough_circles : enable to set min_distance_between_circles
- hough_circles : add debug_image_publisher
- hough_circles : fix bugs on createTrackver uses gaussian_blur_size for sigma x/y
- Contributors: Kei Okada, Iori Yanokura
- Add parameter to people_detector #9
- hough_circles: enable to set double value to the HoughCircle params #8
- hough_circle enable to set gaussian_blue_size and kernel sigma from cfg
- hough_circles: fix default/min/max value of cfg
- hough_circle: enable to set db to 100
- circle_hough: dp, accumrate_threshold, canny_threshold is double, not int
- Add parameter to hough_circles_nodelet #7
- Add parameter to hough_lines_nodelet #6
- Add parameter to edge_detection_nodelet(canny) #5
- Simplify source tree by removing duplicated node codes #4 Closes #3
- fix .travis file
- copy Travis and .gitignore from vision_opencv
- geometry_msgs doesn't get used by opencv_apps, but std_msgs does. (#119)
- Contributors: Kei Okada, Kentaro Wada, Lucas Walter, Vincent Rabaud, IorI Yanokura
- relax test condition
- fix test hz to 5 hz, tested on core i7 3.2G
- Refactor opencv_apps to remove duplicated codes to handle connection of topics. 1. Add opencv_apps::Nodelet class to handle connection and disconnection of topics. 2. Update nodelets of opencv_apps to inhereit opencv_apps::Nodelet class to remove duplicated codes.
- Contributors: Kei Okada, Ryohei Ueda
- check if opencv_contrib is available
- Use respawn instead of watch
- Contributors: Kei Okada, trainman419
- enable simple_flow on opencv3, https://github.com/ros-perception/vision_opencv/commit/8ed5ff5c48b4c3d270cd8216175cf6a8441cb046 can revert https://github.com/ros-perception/vision_opencv/commit/89a933aef7c11acdb75a17c46bfcb60389b25baa
- lk_flow_nodeletcpp, fback_flow_nodelet.cpp: need to copy input image to gray
- opencv_apps: add test programs, this will generate images for wiki
- fix OpenCV3 build
- phase_corr: fix display, bigger circle and line
- goodfeature_track_nodelet.cpp: publish good feature points as corners
- set image encoding to bgr8
- convex_hull: draw hull with width 4
- watershed_segmentatoin_nodelet.cpp: output segmented area as contours and suppot add_seed_points as input of segmentatoin seed
- src/nodelet/segment_objects_nodelet.cpp: change output image topic name from segmented to image
- Convert rgb image to bgr in lk_flow
- [oepncv_apps] fix bug for segment_objects_nodelet.cpp
- Contributors: Kei Okada, Kentaro Wada, Shingo Kitagawa, Vincent Rabaud
- Accept grayscale images as input as well
- Add format enum for easy use and choose format.
- Contributors: Felix Mauch, talregev
- simplify the OpenCV3 compatibility
- fix image output of fback_flow
- fix error: ISO C++ forbids initialization of member for gcc 4.6
- add std_srvs
- add std_srvs
- fix error: ISO C++ forbids initialization of member for gcc 4.6
- get opencv_apps to compile with OpenCV3
- fix licenses for Kei
- add opencv_apps, proposed in #40
- Contributors: Kei Okada, Vincent Rabaud, Yuto Inagaki