Hi,

I'm trying to pull a video feed from rviz into OpenCV to do some real-time image processing on it. For this I publish the rviz video feed onto a ROS topic: /camerawebcam/rgb/imageraw

I'm using:

• Ubuntu 14.04
• ROS Indigo
• OpenCV 3.1
• Rviz 1.11.14

I wrote a ROS node, named talker, to pull in this video feed an do some processing on it. When using the capture function in combination with a webcam feed or a pre-recorded video everything works fine. But now I need to convert the ROS image fro mthe ROS topic to a Mat in order to be able to process it. To do this I wrote a imageCallback function containing the cv_bridge::toCvCopy() function one requires to do this.

All relevant code is listed below.

CMakelist.txt:

cmake_minimum_required(VERSION 2.8.3)
project(camera_corridor_ROS_OpenCV)

  find_package(
  catkin REQUIRED COMPONENTS 
  roscpp 
  std_msgs   
  rospy
  cv_bridge
  image_transport
  message_generation)

# don't use opencv as found by any prior running of find_package
unset(OpenCV_CONFIG_PATH CACHE)
unset(OpenCV_DIR CACHE)

set(TMP_PREFIX_PATH ${CMAKE_PREFIX_PATH})
set(CMAKE_PREFIX_PATH "$ENV{HOME}/opencv-3.1.0")
find_package(OpenCV 3.0 REQUIRED)

# restore CMAKE_PREFIX_PATH and other cached variables
# so nothing other package finds this opencv
set(CMAKE_PREFIX_PATH ${TMP_PREFIX_PATH})

unset(OpenCV_CONFIG_PATH CACHE)
unset(OpenCV_DIR CACHE)

add_message_files(
  FILES
  Corridor_vanish.msg
)
generate_messages(
  DEPENDENCIES
  std_msgs  # Or other packages containing msgs
)
catkin_package(
  CATKIN_DEPENDS message_runtime std_msgs cv_bridge image_transport
)
include_directories(${catkin_INCLUDE_DIRS} include ${OpenCV_INCLUDE_DIRS})
add_library(talker2 src/KalmanDHU.cpp)
add_executable(talker src/talker.cpp)
target_link_libraries(talker ${catkin_LIBRARIES})
target_link_libraries(talker ${OpenCV_LIBRARIES})
target_link_libraries(talker2 ${OpenCV_LIBRARIES})
target_link_libraries(talker talker2)

Package.xml:

<?xml version="1.0"?>
<package>
  <name>camera_corridor_ROS_OpenCV</name>
  <version>0.0.0</version>
  <description>The camera_corridor_ROS_OpenCV package</description>

  <buildtool_depend>catkin</buildtool_depend>
  <build_depend>image_transport</build_depend>
  <build_depend>roscpp</build_depend>
  <build_depend>std_msgs</build_depend>
  <build_depend>cv_bridge</build_depend>
  <build_depend>uvc_camera</build_depend>
  <build_depend>sensor_msgs</build_depend>
  <build_depend>message_generation</build_depend>
  <build_depend>rospy</build_depend> 

  <run_depend>image_transport</run_depend>
  <run_depend>roscpp</run_depend>
  <run_depend>std_msgs</run_depend>
  <run_depend>cv_bridge</run_depend>
  <run_depend>uvc_camera</run_depend>
  <run_depend>sensor_msgs</run_depend>
  <run_depend>message_runtime</run_depend>

  <export>
  </export>
</package>

My main code:

//Includes all the headers necessary to use the most common public pieces of the ROS system.
#include <ros/ros.h>
//Use image_transport for publishing and subscribing to images in ROS
#include <image_transport/image_transport.h>
//Use cv_bridge to convert between ROS and OpenCV Image formats
#include <cv_bridge/cv_bridge.h>
//Include some useful constants for image encoding. Refer to: http://www.ros.org/doc/api/sensor_msgs/html/namespacesensor__msgs_1_1image__encodings.html for more info.
#include <sensor_msgs/image_encodings.h>
//Include headers for OpenCV Image processing
#include <opencv2/imgproc/imgproc.hpp>
//Include headers for OpenCV GUI handling
#include <opencv2/highgui/highgui.hpp>
#include <opencv2/opencv.hpp>


#include "std_msgs/String.h"

#include <sstream>

#include <iostream>
#include "KalmanDHU.h"

#include <camera_corridor_ROS_OpenCV/Corridor_vanish.h>

//Store all constants for image encodings in the enc namespace to be used later.
namespace enc = sensor_msgs::image_encodings;

using namespace std;
using namespace cv;

void imageCallback(const sensor_msgs::ImageConstPtr& original_image);
void ProcessFrame(Mat inputframe);
bool intersection(Point2f o1, Point2f p1, Point2f o2, Point2f p2,Point2f &r);

Point2f pointVanish,pointLeft,pointRight,pointMiddle;
bool corridorFound = false;
vector<Point2f> intersectvector;
int framecounter = 0;
int framesgood = 0;
int frame_last_good=0;

float theta_lim_1 = 0.35;
float theta_lim_2 = 1.20;
float theta_lim_3 = 1.90;
float theta_lim_4 = 2.70;

cv_bridge::CvImagePtr cv_ptr;
Mat source_img;

/**
The code processes a corridor real-time video feed to find perspective lines and a vanishing point for navigational purposes
*/
int main(int argc, char **argv)
{

        ros::init(argc, argv, "image_processor");

        ros::NodeHandle nh;

            ros::Subscriber chatter_sub = nh.subscribe("/camera_webcam/rgb/image_raw", 1, imageCallback);

        ros::Publisher chatter_pub = nh.advertise<camera_corridor_ROS_OpenCV::Corridor_vanish>("corridor_vanish", 10);
            ros::Rate loop_rate(15);

    //Export frames individually
    Mat frame;

    frame = source_img.clone();

    while (ros::ok())
    {

       IMAGE PROCESSING DONE HERE

                chatter_pub.publish(msg);

       }
            ros::spinOnce();

            loop_rate.sleep();
    }

    return 0;   
}

The imageCallback function:

void imageCallback(const sensor_msgs::ImageConstPtr& original_image)
{
    try
    {
            //Always copy, returning a mutable CvImage
            //OpenCV expects color images to use BGR channel order.
            cv_ptr = cv_bridge::toCvCopy(original_image, sensor_msgs::image_encodings::BGR8);
    }
    catch (cv_bridge::Exception& e)
    {
            //if there is an error during conversion, display it
            ROS_ERROR("obstacle_avoidance::main.cpp::cv_bridge exception: %s", e.what());
            return;
    }
    source_img = cv_ptr->image;
}

Compilation goes well. But when I try to run the ROS node talker containing this code I get an instant segmentation fault without any further information:

jon@jon-werk ~          $ rosrun --debug camera_corridor_ROS_OpenCV talker 
[rosrun] Looking in catkin libexec dirs: /home/jon/indigo_ws/devel/lib/camera_corridor_ROS_OpenCV
[rosrun] Looking in rospack dir: /home/jon/indigo_ws/src/itasc_quad_collision/camera_corridor_ROS_OpenCV
[rosrun] Searching for talker with permissions /111
[rosrun] Running  /home/jon/indigo_ws/devel/lib/camera_corridor_ROS_OpenCV/talker
Segmentation fault (core dumped)

I don't know the reason for this. I've read on other fora that it might be due to me using OpenCV 3.1 while cv_bridge for ROS Indigo was intended for OpenCv 2.4, but I'm not sure about this.

Any ideas?

Kind regards,

Jon