【征集/转载】角蜂鸟AI视觉套件4.ROS下订阅并处理图像

feixiang20

[征集]你玩AI，我送幸运，为AI套件专区添砖加瓦--4角蜂鸟AI视觉套件：（四）ROS下订阅并处理图像

对于熟悉ROS的朋友们来说，图像的topic有了，我们就可以开始自己想干的事情了。这里我创建一个名为hs_image_sub的package来处理角蜂鸟的图像，图像的topic 名字为上一篇提到的：/hs/camera/image_raw. 也是我们打开rqt_image_view窗口看到的东西。

这篇博客主要是利用opencv来获取角蜂鸟图像，然后做一个阈值处理，并且打开窗口显示原图和处理后的图像。有了opencv想做什么处理都可以了，这篇文章主要是教大家在ROS里获取并调用图像。

cd ~/catkin_ws/src

catkin_create_pkg hs_image_sub roscpp sensor_msgs cv_bridge

然后

cd ~/catkin_ws/src/hs_image_sub/src

touch hs_image_sub_node.cpp

gedit hs_image_sub_node.cpp

填入以下代码

#include <ros/ros.h>

#include <sensor_msgs/Image.h>

#include <sensor_msgs/image_encodings.h>

#include <image_transport/image_transport.h>

#include <cv_bridge/cv_bridge.h>

// OpenCV

#include <opencv2/opencv.hpp>

#include <opencv2/imgproc/imgproc.hpp>

#include <opencv2/highgui/highgui.hpp>

using namespace std;

using namespace cv;

const string Original_winName = "Original Image";

const string Thresh_winName = "Threshed Image";

Mat cameraFeed;

Mat HSV;

Mat threshold_ori;

void rgbCallback(const sensor_msgs::ImageConstPtr& msg)

{

    cv_bridge::CvImageConstPtr cv_ptr;

    try

    {

        cv_ptr = cv_bridge::toCvCopy(msg, sensor_msgs::image_encodings::BGR8); // Caution the type here.

    }

    catch (cv_bridge::Exception& ex)

    {

        ROS_ERROR("cv_bridge exception in rgbcallback: %s", ex.what());

        exit(-1);

    }

    cameraFeed = cv_ptr->image.clone();

    cvtColor(cameraFeed,HSV,COLOR_BGR2HSV);

    inRange(HSV,Scalar(0,133,0),Scalar(21,256,256),threshold_ori);

   

    //show frames

    imshow(Original_winName,cameraFeed);

    imshow(Thresh_winName,threshold_ori);

    //delay 10ms so that screen can refresh.

    //image will not appear without this waitKey() command

    waitKey(10);

}

int main(int argc, char *argv[])

{

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

    ros::NodeHandle n;

//        topic name of HornedSungem

    ros::Subscriber rgb_sub = n.subscribe("/hs/camera/image_raw", 1, rgbCallback);

    ROS_INFO("Subscribe to the HS color image topic.");

    ros::spin();

    return 0;

}

代码比较简单，主要是订阅图像源，调用回调函数做阈值处理。

注意这里的topic name 和回调函数里的格式：BGR8，只能是这个。

保存，然后

cd ~/catkin_ws/src/hs_image_sub/

gedit CMakeLists.txt

全选，替换为以下内容

cmake_minimum_required(VERSION 2.8.3)

project(hs_image_sub)

find_package(OpenCV REQUIRED)

find_package(catkin REQUIRED COMPONENTS

  cv_bridge

  roscpp

  sensor_msgs

)

catkin_package(

#  INCLUDE_DIRS include

#  LIBRARIES hs_image_sub

#  CATKIN_DEPENDS cv_bridge roscpp sensor_msgs

#  DEPENDS system_lib

)

include_directories(

# include

  ${catkin_INCLUDE_DIRS}

)

add_executable(${PROJECT_NAME}_node src/hs_image_sub_node.cpp)

target_link_libraries(${PROJECT_NAME}_node

   ${catkin_LIBRARIES} ${OpenCV_LIBS}

)

接着编译

cd ~/catkin_ws

catkin_make

看到编译成功，运行角蜂鸟ROS的官方程序，启动相机发布图像源，来供我们这里的程序调用

roslaunch horned_sungem_launch hs_camera.launch cnn_type:=googlenet camera:=hs pixels:=360
rosrun hs_image_sub hs_image_sub_node

有了opencv想做什么处理都可以了比如追踪

版权声明：本文为博主原创文章，如需转载，请附上文章原文地址。        https://blog.csdn.net/yaked/article/details/81477833

satoll

代码部分排版有问题，请查实

feixiang20

satoll 发表于 2018-8-20 09:48
代码部分排版有问题，请查实

改好了