Open vicFigure opened 7 years ago
I have the same prolbem, so I change the code in line 89 to:
const op::Point<int> netInputSize_Point(netInputSize.height, netInputSize.width);
openPoseCvMatToOpInput = std::shared_ptr<op::CvMatToOpInput>(
new op::CvMatToOpInput(netInputSize_Point, numScales, (float)scaleGap)
);
Also the other errors caused by "cv::Size" and "op::Point" are fixed in the same way.
Then I can successfully compile it. But when I run rosrun openpose-ros openpose-ros-node, I got a Segment fault error. @vicFigure
Hi, I am facing the same issue as well. Has anyone managed to make progress on this issue?
@QichaoXu Facing the same issue, Were you able to fix the seg fault?
Issure Summary
When I in my catkin_workspace, and run catkin_make, some error occurs like this:
/home/vic/catkin_ws_test/src/openpose-ros/src/openpose_ros_node.cpp:91:76: error: no matching function for call to ‘op::CvMatToOpInput::CvMatToOpInput(cv::Size&, int&, float)’ new op::CvMatToOpInput(netInputSize, numScales, (float)scaleGap) ^ In file included from /home/vic/openpose/include/openpose/core/headers.hpp:7:0, from /home/vic/catkin_ws_test/src/openpose-ros/src/openpose_ros_node.cpp:5: /home/vic/openpose/include/openpose/core/cvMatToOpInput.hpp:13:9: note: candidate: op::CvMatToOpInput::CvMatToOpInput(const op::Point<int>&, int, float) CvMatToOpInput(const Point<int>& netInputResolution, const int scaleNum ^ /home/vic/openpose/include/openpose/core/cvMatToOpInput.hpp:13:9: note: no known conversion for argument 1 from ‘cv::Size {aka cv::Size_<int>}’ to ‘const op::Point<int>&’ /home/vic/openpose/include/openpose/core/cvMatToOpInput.hpp:10:18: note: candidate: op::CvMatToOpInput::CvMatToOpInput(const op::CvMatToOpInput&) class OP_API CvMatToOpInput ^ /home/vic/openpose/include/openpose/core/cvMatToOpInput.hpp:10:18: note: candidate expects 1 argument, 3 provided /home/vic/openpose/include/openpose/core/cvMatToOpInput.hpp:10:18: note: candidate: op::CvMatToOpInput::CvMatToOpInput(op::CvMatToOpInput&&) /home/vic/openpose/include/openpose/core/cvMatToOpInput.hpp:10:18: note: candidate expects 1 argument, 3 provided
There are also some other erros, saying that 1)."no matching function" in the cpp files of openpose,2)."no known conversion for argument 1 from ‘cv::Size {aka cv::Size_
I don't know what's wrong about my openpose or openpose-ros? Can somebody help me? Thank you very much.
System Configuration
Ubuntu 16.04 cuda 8.0 cudnn 6.0 Opencv3.2.0