请用c++写一段ros程序，gazebo仿真环境下有一个realsense深度相机,要求读取深度相机的深度图像和rgb图像，两者的话题名分别为/camera3/depth/image_raw和/camera3/rgb/image_raw，深度图片的编码格式为32FC1，读取图片并分别显示

好的，这是一个简单的ROS程序，可以读取realsense深度相机的深度图像和rgb图像，并在rviz中分别显示。

#include <ros/ros.h>
#include <sensor_msgs/Image.h>
#include <sensor_msgs/image_encodings.h>
#include <cv_bridge/cv_bridge.h>
#include <opencv2/highgui/highgui.hpp>

// 定义全局变量，用于存储深度图像和RGB图像
cv::Mat depth_image;
cv::Mat rgb_image;

// 深度图像的回调函数
void depthCallback(const sensor_msgs::ImageConstPtr& msg)
{
  // 将ROS消息转换为OpenCV图像
  cv_bridge::CvImagePtr cv_ptr;
  try
  {
    cv_ptr = cv_bridge::toCvCopy(msg, sensor_msgs::image_encodings::TYPE_32FC1);
  }
  catch (cv_bridge::Exception& e)
  {
    ROS_ERROR("cv_bridge exception: %s", e.what());
    return;
  }

  // 保存深度图像
  depth_image = cv_ptr->image;
}

// RGB图像的回调函数
void rgbCallback(const sensor_msgs::ImageConstPtr& msg)
{
  // 将ROS消息转换为OpenCV图像
  cv_bridge::CvImagePtr cv_ptr;
  try
  {
    cv_ptr = cv_bridge::toCvCopy(msg, sensor_msgs::image_encodings::BGR8);
  }
  catch (cv_bridge::Exception& e)
  {
    ROS_ERROR("cv_bridge exception: %s", e.what());
    return;
  }

  // 保存RGB图像
  rgb_image = cv_ptr->image;
}

int main(int argc, char** argv)
{
  // 初始化ROS节点
  ros::init(argc, argv, "realsense_reader");
  ros::NodeHandle nh;

  // 订阅深度图像和RGB图像的话题
  ros::Subscriber depth_sub = nh.subscribe("/camera3/depth/image_raw", 1, depthCallback);
  ros::Subscriber rgb_sub = nh.subscribe("/camera3/rgb/image_raw", 1, rgbCallback);

  // 创建OpenCV窗口
  cv::namedWindow("Depth Image");
  cv::namedWindow("RGB Image");

  // 循环读取深度图像和RGB图像，并在OpenCV窗口中显示
  while (ros::ok())
  {
    // 显示深度图像
    if (!depth_image.empty())
    {
      cv::imshow("Depth Image", depth_image);
    }

    // 显示RGB图像
    if (!rgb_image.empty())
    {
      cv::imshow("RGB Image", rgb_image);
    }

    // 等待按键，退出程序
    char key = cv::waitKey(10);
    if (key == 27) // ESC键
    {
      break;
    }

    // 处理ROS回调函数
    ros::spinOnce();
  }

  // 销毁OpenCV窗口
  cv::destroyAllWindows();

  return 0;
}

需要注意的是，该程序中使用了OpenCV库来显示深度图像和RGB图像，因此需要在CMakeLists.txt文件中添加OpenCV的依赖项。完整的CMakeLists.txt文件如下所示：

cmake_minimum_required(VERSION 2.8.3)
project(realsense_reader)

find_package(catkin REQUIRED COMPONENTS
  roscpp
  sensor_msgs
  cv_bridge
  image_transport
)

find_package(OpenCV REQUIRED)

catkin_package(
  CATKIN_DEPENDS roscpp sensor_msgs cv_bridge image_transport
)

include_directories(
  ${catkin_INCLUDE_DIRS}
  ${OpenCV_INCLUDE_DIRS}
)

add_executable(${PROJECT_NAME} src/realsense_reader.cpp)

target_link_libraries(${PROJECT_NAME}
  ${catkin_LIBRARIES}
  ${OpenCV_LIBRARIES}
)

install(TARGETS ${PROJECT_NAME}
  RUNTIME DESTINATION ${CATKIN_PACKAGE_BIN_DESTINATION}
)

另外，需要在.launch文件中启动Gazebo仿真环境和realsense深度相机节点，并将/camera3/depth/image_raw和/camera3/rgb/image_raw话题分别映射到程序中定义的回调函数中。完整的.launch文件如下所示：

<launch>
  <!-- 启动Gazebo仿真环境 -->
  <include file="$(find gazebo_ros)/launch/empty_world.launch">
    <arg name="world_name" value="$(find realsense_gazebo)/worlds/realsense_world.world"/>
  </include>

  <!-- 启动realsense深度相机节点 -->
  <node name="realsense_node" pkg="realsense2_camera" type="realsense2_camera_node">
    <param name="serial_no" value="123456789"/>
    <param name="enable_depth" value="true"/>
    <param name="enable_color" value="true"/>
    <param name="depth_height" value="480"/>
    <param name="depth_width" value="640"/>
    <param name="color_height" value="480"/>
    <param name="color_width" value="640"/>
    <param name="depth_fps" value="30"/>
    <param name="color_fps" value="30"/>
  </node>

  <!-- 映射/camera3/depth/image_raw话题 -->
  <node name="depth_image" pkg="image_transport" type="republish">
    <remap from="/image_raw" to="/camera3/depth/image_raw"/>
    <remap from="/compressed" to="/camera3/depth/compressed"/>
    <remap from="/compressedDepth" to="/camera3/depth/compressedDepth"/>
    <remap from="/theora" to="/camera3/depth/theora"/>
  </node>

  <!-- 映射/camera3/rgb/image_raw话题 -->
  <node name="rgb_image" pkg="image_transport" type="republish">
    <remap from="/image_raw" to="/camera3/rgb/image_raw"/>
    <remap from="/compressed" to="/camera3/rgb/compressed"/>
    <remap from="/compressedDepth" to="/camera3/rgb/compressedDepth"/>
    <remap from="/theora" to="/camera3/rgb/theora"/>
  </node>

  <!-- 启动realsense_reader节点 -->
  <node name="realsense_reader" pkg="realsense_reader" type="realsense_reader"/>
</launch>

启动.launch文件后，即可在rviz中看到深度图像和RGB图像的显示。