Ask Your Question
0

Combines message for image and information gives "invalid initialization error"

asked 2019-04-29 12:45:36 -0600

MarviB16 gravatar image

Hello,

I am currently trying to implement a message that includes the images as well as the detected bounding boxes. The publisher works very well, but the subscriber gives me the following error message:

In file included from /usr/include/boost/function/detail/maybe_include.hpp:18:0,
                 from /usr/include/boost/function/detail/function_iterate.hpp:14,
                 from /usr/include/boost/preprocessor/iteration/detail/iter/forward1.hpp:52,
                 from /usr/include/boost/function.hpp:64,
                 from /opt/ros/kinetic/include/ros/forwards.h:40,
                 from /opt/ros/kinetic/include/ros/common.h:37,
                 from /opt/ros/kinetic/include/ros/ros.h:43,
                 from /home/marvin/catkin_ws/src/efr_object_detection_camera/src/template_matching.cpp:5:
/usr/include/boost/function/function_template.hpp: In instantiation of ‘static void boost::detail::function::void_function_invoker1<FunctionPtr, R, T0>::invoke(boost::detail::function::function_buffer&, T0) [with FunctionPtr = void (*)(const darknet_ros_msgs::ImageWithBoundingBoxes_<std::allocator<void> >&); R = void; T0 = const boost::shared_ptr<const sensor_msgs::Image_<std::allocator<void> > >&]’:
/usr/include/boost/function/function_template.hpp:940:38:   required from ‘void boost::function1<R, T1>::assign_to(Functor) [with Functor = void (*)(const darknet_ros_msgs::ImageWithBoundingBoxes_<std::allocator<void> >&); R = void; T0 = const boost::shared_ptr<const sensor_msgs::Image_<std::allocator<void> > >&]’
/usr/include/boost/function/function_template.hpp:728:7:   required from ‘boost::function1<R, T1>::function1(Functor, typename boost::enable_if_c<boost::type_traits::ice_not<boost::is_integral<Functor>::value>::value, int>::type) [with Functor = void (*)(const darknet_ros_msgs::ImageWithBoundingBoxes_<std::allocator<void> >&); R = void; T0 = const boost::shared_ptr<const sensor_msgs::Image_<std::allocator<void> > >&; typename boost::enable_if_c<boost::type_traits::ice_not<boost::is_integral<Functor>::value>::value, int>::type = int]’
/usr/include/boost/function/function_template.hpp:1077:16:   required from ‘boost::function<R(T0)>::function(Functor, typename boost::enable_if_c<boost::type_traits::ice_not<boost::is_integral<Functor>::value>::value, int>::type) [with Functor = void (*)(const darknet_ros_msgs::ImageWithBoundingBoxes_<std::allocator<void> >&); R = void; T0 = const boost::shared_ptr<const sensor_msgs::Image_<std::allocator<void> > >&; typename boost::enable_if_c<boost::type_traits::ice_not<boost::is_integral<Functor>::value>::value, int>::type = int]’
/home/marvin/catkin_ws/src/efr_object_detection_camera/src/template_matching.cpp:55:26:   required from here
/usr/include/boost/function/function_template.hpp:118:11: error: invalid initialization of reference of type ‘const darknet_ros_msgs::ImageWithBoundingBoxes_<std::allocator<void> >&’ from expression of type ‘const boost::shared_ptr<const sensor_msgs::Image_<std::allocator<void> > >’
           BOOST_FUNCTION_RETURN(f(BOOST_FUNCTION_ARGS));
           ^

This is the message in question:

Header header

Header image_header

BoundingBox[] bounding_boxes

sensor_msgs/Image im

And this is my subscriber:

void imageCallback(const darknet_ros_msgs::ImageWithBoundingBoxes& msg) {
    cv_bridge::CvImagePtr cv_ptr;
    try {
        cv_ptr = cv_bridge::toCvCopy(msg.im, sensor_msgs::image_encodings::BGR8);
        Mat image = cv_ptr->image;
        for(size_t i = 0; i < msg.bounding_boxes.size(); i++)
        {
            darknet_ros_msgs::BoundingBox bBox= msg.bounding_boxes[i];
            double minx = bBox.xmin;
            double miny = bBox.ymin;
            double maxx = bBox.xmax;
            double maxy = bBox.ymax;
            match(image);
        }

        cv::waitKey(30);
    } catch (cv_bridge::Exception& e) {
        ROS_ERROR("Could not convert from '%s' to 'bgr8'.",
        msg.im.encoding.c_str());
    }
}

int main(int argc, char **argv) {
    ros::init(argc, argv, "template_matching_node");
    ros::NodeHandle nh;
    cv::namedWindow("view");
    cv::startWindowThread();
    image_transport::ImageTransport it(nh);
    image_transport::Subscriber sub = it.subscribe("/darknet_ros/ImageWithBoundingBoxes", 1,
            imageCallback);
    ros::spin ...
(more)
edit retag flag offensive close merge delete

1 Answer

Sort by » oldest newest most voted
1

answered 2019-04-29 13:14:07 -0600

gvdhoorn gravatar image

updated 2019-04-29 13:41:26 -0600

Isn't the problem here that you're trying to use image_transport::Subscriber to subscribe to a topic which isn't actually carrying Images, but darknet_ros_msgs::ImageWithBoundingBoxes?

That's also what the error message seems to tell you. Paraphrasing, it would read:

error: invalid initialization of reference of type ‘darknet_ros_msgs::ImageWithBoundingBoxes&’ from expression of type ‘boost::shared_ptr<const sensor_msgs::Image>

You'll probably have to switch to a regular ros::Subscriber instead of using image_transport's.


Edit:

Oh, right yes... But how could I rewrite it? So that I still can use the image_transport functionality?

I'm not sure. Possibly not using a custom message, but using something like message_filters to combine related messages based on their timestamps (so only publish a list of boundingboxes (one from vision_msgs preferably) with the same timestamp as the sensor_msgs/Image that the detection was run on, then later recombine the two using message_filters).

I'm not an image_transport expert though, so there could be a different option.

There is a tutorial on writing your own transport plugin (this one), but I don't believe that is something that would fit your use-case.


Edit 2:

Edit: Thank you! Now it works. But what I am still not sure about: Is it now slower? Because i don't use the image transport functionality?

Using image_transport doesn't make anything magically faster. The "only" thing image_transport does is making it easier to use different compression algorithms (called "transports" in image_transport lingo) without having to deal with their details in image producers and consumers. Simply configure a specific transport to be used, and without recompiling or changing anything in your publishers or subscribers it will be used. Transparently.

In the case of compression, this actually has the potential to make things slower (as compression adds overhead), but that obviously depends on a number of other factors (size of images vs bandwidth, latency, etc).

So with switching to a regular Subscriber, you lose all of that. But you had already lost that, as you were using a regular Publisher as well.

edit flag offensive delete link more

Comments

Oh, right yes... But how could I rewrite it? So that I still can use the image_transport functionality?

Edit: Thank you! Now it works. But what I am still not sure about: Is it now slower? Because i don't use the image transport functionality?

MarviB16 gravatar imageMarviB16 ( 2019-04-29 13:21:47 -0600 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Question Tools

2 followers

Stats

Asked: 2019-04-29 12:45:36 -0600

Seen: 239 times

Last updated: Apr 29 '19