ROS下利用realsense采集RGBD图像合成点云

摘要：在ROS kinetic下，利用realsense D435深度相机采集校准的RGBD图片，合成点云，在rviz中查看点云，最后保存成pcd文件。

一、各种bug

代码编译成功后，打开rviz添加pointcloud2选项卡，当我订阅合成点云时，可视化失败，选项卡报错：

1）Data size (9394656 bytes) does not match width (640) times height (480) times point_step (32). Dropping message.

解读：通过 rostopic echo /pointcloud_topic 读取相机节点发布的原始点云的相关数据，可以发现每一帧原始点云的点数量为307200。合成点云的点数量为 9394656 / 32 ,约26万。可以猜测由于某种原因，系统把每一帧合成点云的数据都丢弃了。

原因：我事先给定合成点云的大小为，height = 480，width = 640. 然而在合成点云的过程中，剔除了部分违法值(d=0)，由此导致合成点云的点数量与指定的点数量不匹配，合成点云的数据因此被丢弃。

解决方法：采用如下方法给定点云大小， cloud->height = 1; cloud->width = cloud->points.size();

2）transform xxxxx；

解读：通过 rostopic echo /pointcloud_topic ，发现原始点云数据具有如下信息，

header: 
  seq: 50114
  stamp: 
    secs: 1528439158
    nsecs: 557543003
  frame_id: "camera_color_optical_frame"

由此推断，合成点云缺失参考坐标系header.frame_id。点云中点的XYZ属性是相对于某个坐标系来描述的，因此，需要指定点云的参考坐标系。

解决方法：添加点云的header信息，

pub_pointcloud.header.frame_id = "camera_color_optical_frame"; 
pub_pointcloud.header.stamp = ros::Time::now();

3）将合成的点云存储成pcd文件时遇到如下报错：

[ INFO] [1528442016.931874649]: point cloud size = 0
terminate called after throwing an instance of 'pcl::IOException'
  what():  : [pcl::PCDWriter::writeASCII] Input point cloud has no data!
Aborted (core dumped)

经过多方查找，摸索了一步trick，分享给大家。真实报错原因仍未查明，期待前辈的指点。

高博的源代码如下所示，里面的cloud是pcl的数据类型，

pcl::io::savePCDFile( "./pointcloud.pcd", *cloud ); 。

我的源代码如下面所示，先通过 pcl::toROSMsg() 将pcl的数据类型转换成ros的数据类型，再写入pcd中，即可跳过报错。

4）相机内参

由于在彩色图和深度图配准的过程中，选用的是彩色图的坐标系，因此在合成点云（像素坐标在变换到相机坐标）时应该选用彩色图的相机内参。

realsense官方提供了一个应用程序可以查看所有视频流的内参。

gordon@gordon-5577:/usr/local/bin$ ./rs-sensor-control

如下所示

84 : Color #0 (Video Stream: Y16 640x480@ 60Hz)
85 : Color #0 (Video Stream: BGRA8 640x480@ 60Hz)
86 : Color #0 (Video Stream: RGBA8 640x480@ 60Hz)
87 : Color #0 (Video Stream: BGR8 640x480@ 60Hz)
88 : Color #0 (Video Stream: RGB8 640x480@ 60Hz)
89 : Color #0 (Video Stream: YUYV 640x480@ 60Hz)
90 : Color #0 (Video Stream: Y16 640x480@ 30Hz)
91 : Color #0 (Video Stream: BGRA8 640x480@ 30Hz)
92 : Color #0 (Video Stream: RGBA8 640x480@ 30Hz)
93 : Color #0 (Video Stream: BGR8 640x480@ 30Hz)
94 : Color #0 (Video Stream: RGB8 640x480@ 30Hz)
95 : Color #0 (Video Stream: YUYV 640x480@ 30Hz)
96 : Color #0 (Video Stream: Y16 640x480@ 15Hz)
97 : Color #0 (Video Stream: BGRA8 640x480@ 15Hz)
98 : Color #0 (Video Stream: RGBA8 640x480@ 15Hz)
99 : Color #0 (Video Stream: BGR8 640x480@ 15Hz)
100: Color #0 (Video Stream: RGB8 640x480@ 15Hz)
101: Color #0 (Video Stream: YUYV 640x480@ 15Hz)
102: Color #0 (Video Stream: Y16 640x480@ 6Hz)
103: Color #0 (Video Stream: BGRA8 640x480@ 6Hz)
104: Color #0 (Video Stream: RGBA8 640x480@ 6Hz)
105: Color #0 (Video Stream: BGR8 640x480@ 6Hz)
106: Color #0 (Video Stream: RGB8 640x480@ 6Hz)
107: Color #0 (Video Stream: YUYV 640x480@ 6Hz)

5）深度图从ROS的数据类型转换为CV的数据类型

参看链接

二、程序代码

#include <ros/ros.h>
#include <image_transport/image_transport.h>
#include <opencv2/highgui/highgui.hpp>
#include <cv_bridge/cv_bridge.h>
#include <sensor_msgs/image_encodings.h>
#include <sensor_msgs/PointCloud2.h>

// PCL 库
#include <pcl/io/pcd_io.h>
#include <pcl/point_types.h>
#include <pcl/visualization/pcl_visualizer.h>
#include <pcl_conversions/pcl_conversions.h>
 
// 定义点云类型
typedef pcl::PointCloud<pcl::PointXYZRGB> PointCloud; 

using namespace std;
//namespace enc = sensor_msgs::image_encodings;

// 相机内参
const double camera_factor = 1000;
const double camera_cx = 321.798;
const double camera_cy = 239.607;
const double camera_fx = 615.899;
const double camera_fy = 616.468;

// 全局变量：图像矩阵和点云
cv_bridge::CvImagePtr color_ptr, depth_ptr;
cv::Mat color_pic, depth_pic;

void color_Callback(const sensor_msgs::ImageConstPtr& color_msg)
{
  //cv_bridge::CvImagePtr color_ptr;
  try
  {
    cv::imshow("color_view", cv_bridge::toCvShare(color_msg, sensor_msgs::image_encodings::BGR8)->image);
    color_ptr = cv_bridge::toCvCopy(color_msg, sensor_msgs::image_encodings::BGR8);    

    cv::waitKey(1050); // 不断刷新图像，频率时间为int delay，单位为ms
  }
  catch (cv_bridge::Exception& e )
  {
    ROS_ERROR("Could not convert from '%s' to 'bgr8'.", color_msg->encoding.c_str());
  }
  color_pic = color_ptr->image;

  // output some info about the rgb image in cv format
  cout<<"output some info about the rgb image in cv format"<<endl;
  cout<<"rows of the rgb image = "<<color_pic.rows<<endl; 
  cout<<"cols of the rgb image = "<<color_pic.cols<<endl; 
  cout<<"type of rgb_pic's element = "<<color_pic.type()<<endl; 
}


void depth_Callback(const sensor_msgs::ImageConstPtr& depth_msg)
{
  //cv_bridge::CvImagePtr depth_ptr;
  try
  {
    //cv::imshow("depth_view", cv_bridge::toCvShare(depth_msg, sensor_msgs::image_encodings::TYPE_16UC1)->image);
    //depth_ptr = cv_bridge::toCvCopy(depth_msg, sensor_msgs::image_encodings::TYPE_16UC1); 
    cv::imshow("depth_view", cv_bridge::toCvShare(depth_msg, sensor_msgs::image_encodings::TYPE_32FC1)->image);
    depth_ptr = cv_bridge::toCvCopy(depth_msg, sensor_msgs::image_encodings::TYPE_32FC1); 

    cv::waitKey(1050);
  }
  catch (cv_bridge::Exception& e)
  {
    ROS_ERROR("Could not convert from '%s' to 'mono16'.", depth_msg->encoding.c_str());
  }

  depth_pic = depth_ptr->image;

  // output some info about the depth image in cv format
  cout<<"output some info about the depth image in cv format"<<endl;
  cout<<"rows of the depth image = "<<depth_pic.rows<<endl; 
  cout<<"cols of the depth image = "<<depth_pic.cols<<endl; 
  cout<<"type of depth_pic's element = "<<depth_pic.type()<<endl; 
}
int main(int argc, char **argv)
{
  ros::init(argc, argv, "image_listener");
  ros::NodeHandle nh;
  cv::namedWindow("color_view");
  cv::namedWindow("depth_view");
  cv::startWindowThread();
  image_transport::ImageTransport it(nh);
  image_transport::Subscriber sub = it.subscribe("/camera/color/image_raw", 1, color_Callback);
  image_transport::Subscriber sub1 = it.subscribe("/camera/aligned_depth_to_color/image_raw", 1, depth_Callback);
  ros::Publisher pointcloud_publisher = nh.advertise<sensor_msgs::PointCloud2>("generated_pc", 1);
 // 点云变量
  // 使用智能指针，创建一个空点云。这种指针用完会自动释放。
  PointCloud::Ptr cloud ( new PointCloud );
  sensor_msgs::PointCloud2 pub_pointcloud;

  double sample_rate = 1.0; // 1HZ，1秒发1次 
  ros::Rate naptime(sample_rate); // use to regulate loop rate 

  cout<<"depth value of depth map : "<<endl;

  while (ros::ok()) {
    // 遍历深度图
    for (int m = 0; m < depth_pic.rows; m++){
      for (int n = 0; n < depth_pic.cols; n++){
          // 获取深度图中(m,n)处的值
          float d = depth_pic.ptr<float>(m)[n];//ushort d = depth_pic.ptr<ushort>(m)[n];
          // d 可能没有值，若如此，跳过此点
          if (d == 0)
             continue;
          // d 存在值，则向点云增加一个点
          pcl::PointXYZRGB p;

          // 计算这个点的空间坐标
          p.z = double(d) / camera_factor;
          p.x = (n - camera_cx) * p.z / camera_fx;
          p.y = (m - camera_cy) * p.z / camera_fy;
            
          // 从rgb图像中获取它的颜色
          // rgb是三通道的BGR格式图，所以按下面的顺序获取颜色
          p.b = color_pic.ptr<uchar>(m)[n*3];
          p.g = color_pic.ptr<uchar>(m)[n*3+1];
          p.r = color_pic.ptr<uchar>(m)[n*3+2];
        
          // 把p加入到点云中
          cloud->points.push_back( p );
      }
    }
        

    // 设置并保存点云
    cloud->height = 1;
    cloud->width = cloud->points.size();
    ROS_INFO("point cloud size = %d",cloud->width);
    cloud->is_dense = false;// 转换点云的数据类型并存储成pcd文件
    pcl::toROSMsg(*cloud,pub_pointcloud);
    pub_pointcloud.header.frame_id = "camera_color_optical_frame";
    pub_pointcloud.header.stamp = ros::Time::now();
    pcl::io::savePCDFile("./pointcloud.pcd", pub_pointcloud);
    cout<<"publish point_cloud height = "<<pub_pointcloud.height<<endl;
    cout<<"publish point_cloud width = "<<pub_pointcloud.width<<endl;

    // 发布合成点云和原始点云
    pointcloud_publisher.publish(pub_pointcloud);
    ori_pointcloud_publisher.publish(cloud_msg);
    
    // 清除数据并退出
    cloud->points.clear();

    ros::spinOnce(); //allow data update from callback; 
    naptime.sleep(); // wait for remainder of specified period; 
  }

  cv::destroyWindow("color_view");
  cv::destroyWindow("depth_view");
}

Android进阶之旅-(NDK实战篇之基础知识)

Andriod设计模式实战

热门文章