rosbag相关

最新推荐文章于 2024-06-08 10:34:31 发布

Alan Lan

最新推荐文章于 2024-06-08 10:34:31 发布

阅读量3.3k

点赞数 4

分类专栏： ROS

本文链接：https://blog.csdn.net/A_L_A_N/article/details/84564783

版权

ROS 专栏收录该内容

5 篇文章 0 订阅

订阅专栏

一、命令行

1、录制bag

rosbag record -h # 查看record使用方法
rosbag record -a # 录制所有的topic，将bag文件保存在当前文件夹下，并以录制结束时间(日期加时间)命名
rosbag record /topic1 /topic2 # 录制指定的topic
rosbag record -O <bag_name>.bag /topic # (大写o)指定生成数据包的名字
rosbag record -o <bag_prefix> /topic # (小写o)给数据包的名字加前缀，包名为<bag_prefix>+时间戳.bag
rosbag record -b # buffsize(MB)，Default: 256, 0 = infinite

You could run your rosbag with a Not anonymous node name:
$ rosbag record -o /file/name /topic __name:=my_bag

and then kill it via rosnode kill:
$ rosnode kill /my_bag

This assures that rosbag stops gracefully.

2、回放bag

rosbag play <bag_name> # Hit space to toggle paused, or 's' to step.
rosbag play bag1 bag2 # 回放多个bag
rosbag play <bag_name> --topics /topic1 /topic2 # 回放指定话题
rosbag play <bag_name> /image_raw:=/camera/image_raw # 话题名重映射，将/image_raw映射到/camera/image_raw。另一种方法是在roslaunch中修改，<remap from="/image_raw" to="/camera/image_raw"/>
rosbag play -l <bag_name> # 循环播放
rosbag play -r 0.5 <bag_name> # 以0.5倍的录制频率回放
rosbag play --clock --hz=200 <bag_name> # 指定回放频率，默认100hz
rosbag play <bag_name> -d <delay_time> # 等待一定时间之后发布bag文件中的内容
rosbag play -s 10 <bag_name> # 跳过前面10s，从第10s开始播放
rosbag play --skip-empty=1.0 <bag_name> # 每当遇到超过1秒的间隙时，它会立即及时跳跃
rosbag play --pause <bag_name> # 开始播放立刻暂停，按空格键可以恢复/暂停播放。如果在启动rosbag play之后还有一些其他设置要做，比如在rviz中配置可视化选项等，如果不播放rosbag，而是先设置rviz选项，可能有些topic还没有生成，设置的时候没有办法选择，ros graph里面也不存在对应的topic；如果先播放 rosbag，然后再做其他设置，则可能错过了rosbag最开始的一部分信息。解决方案是在播放rosbag时选择“暂停”功能，这样既可以得到相应的topic，又避免错过开头的信息。

使用仿真时间

rosparam set use_sim_time true
rosbag play <bag_name> --clock

3、消息过滤

rosbag filter src.bag dst.bag "topic == '/foo' or topic == '/bar'" # 根据话题名过滤
rosbag filter src.bag dst.bag "t.to_sec() >= 1573631118.88 and t.to_sec() <= 1573631309.16" # 根据时间戳范围过滤
rosbag filter src.bag dst.bag "topic != 'tf' or ((len(m.transforms) > 0 and m.transforms[0].header.frame_id != 'odom') and (len(m.transforms) <= 1 or (len(m.transforms) > 1 and m.transforms[1].header.frame_id != 'odom')) and (len(m.transforms) <= 2 or (len(m.transforms) > 2 and m.transforms[2].header.frame_id != 'odom')) and (len(m.transforms) <= 3 or (len(m.transforms) > 3 and m.transforms[3].header.frame_id != 'odom')) and (len(m.transforms) <= 4 or (len(m.transforms) > 4 and m.transforms[4].header.frame_id != 'odom')))"

4、查看bag的基本信息

rosbag info -h
rosbag info -y <bag_name> # 输出yaml格式的信息
rosbag info <bag_name> # 显示bag信息（开始和结束时间，包含的话题名称及其类型，消息数量、频率以及压缩统计信息）
rosbag info -y -k duration <bag_name> # 输出bag中指定域的信息，比如只显示持续时间

5、其他相关

# 检查一个bag在当前的系统中是否可以回放
rosbag check <bag_name>

# 如果录制的bag很大，可以进行压缩，默认压缩格式为bz2
rosbag compress <bag_name>
# 手动指定压缩格式为bz2
rosbag compress -j <bag_name>
# 指定压缩格式为LZ4（BZ2比LZ4更小但也更慢）
rosbag compress --lz4 <bag_name>

# 压缩完后需要解压缩
rosbag decompress <bag_name>

# 如果回放遇到问题，提示reindex的话，执行后会自动生成一个原bag的备份
rosbag reindex <bag_name>

# topic重命名
rosrun rosbag topic_renamer.py <in_topic> <in_bag> <out_topic> <out_bag>

# 输出bag中指定topic内容
rostopic echo -b bag_name.bag -p /topic_name

# 压缩图片转化为正常图片消息发布，压缩的图像topic: /camera_obj_center/image_color/compressed
rosrun image_transport republish compressed in:=/camera_obj_center/image_color raw out:=/camera/image_raw

# 使用image_view(或rviz)查看视频流效果
rosrun image_view image_view image:=/image_topic
# 查看压缩视频流:
rosrun image_view image_view image:=/image_basic_topic _image_transport:=compressed

二、C++

1、图像序列生成bag

#include <ros/ros.h>
#include <rosbag/bag.h>
#include <opencv2/highgui/highgui.hpp>
#include "cv_bridge/cv_bridge.h" // opencv图像格式转化为ROS图像格式

int main(int argc, char **argv)
{
    ros::init(argc, argv, "rosbag_recode_node");
    ros::NodeHandle nh;

    rosbag::Bag bag;
    bag.open("/home/alan/Desktop/alan_test.bag", rosbag::bagmode::Write);
    // bag.open("/home/alan/Desktop/alan_test.bag", rosbag::bagmode::Append);
    // bag.setCompression(rosbag::compression::LZ4);

    cv::VideoCapture cap(0);

    while (ros::ok() && (cv::waitKey(1) != 27)) { // 按Esc退出
        cv::Mat frame;
        cap >> frame;
        if (frame.empty()) {
            std::cout << "error, no image" << std::endl;
            return -1;
        }

        cv::imshow("image", frame);
        cv::waitKey(10);

        std_msgs::Header header;
        header.frame_id = "image_frame"; // 设置坐标系，多传感器、RVIZ显示时会用到
        header.stamp = ros::Time::now();

        sensor_msgs::ImagePtr image_msg = cv_bridge::CvImage(header, "bgr8", frame).toImageMsg();
        sensor_msgs::CompressedImagePtr compressed_image_msg = cv_bridge::CvImage(header, "bgr8", frame).toCompressedImageMsg();

        bag.write("image_raw", ros::Time::now(), image_msg); // 当前时间
        bag.write("image_raw/compressed", ros::Time::now(), compressed_image_msg);

        //ros::Time time;
        //time.fromNSec(timestamp); // 从图像数据集生成bag时，timestamp一般为图像名
        //bag.write("image_topic", time, msg);
    }
    bag.close();
    cap.release();
    return 0;
}

CMakeLists.txt:

cmake_minimum_required(VERSION 2.8.3)
project(alan_package)

set(CMAKE_CXX_STANDARD 11)

find_package(OpenCV REQUIRED)
include_directories(${OpenCV_INCLUDE_DIRS})

find_package(catkin REQUIRED COMPONENTS
  roscpp
  rosbag
  cv_bridge
)

include_directories(
  ${catkin_INCLUDE_DIRS}
)

add_executable(create_image_bag src/create_image_bag.cpp)
target_link_libraries(create_image_bag
        ${catkin_LIBRARIES}
        ${OpenCV_LIBS})

说明：

CvBridge中常用的编码：

mono8:  CV_8UC1,  grayscale image
mono16: CV_16UC1, 16-bit grayscale image
bgr8:   CV_8UC3,  color image with blue-green-red color order
rgb8:   CV_8UC3,  color image with red-green-blue color order
bgra8:  CV_8UC4,  BGR color image with an alpha channel
rgba8:  CV_8UC4,  RGB color image with an alpha channel

2、从bag中保存图像

#include <string>
#include <rosbag/bag.h>
#include <rosbag/view.h>
#include <sensor_msgs/Image.h>
#include <sensor_msgs/CompressedImage.h>
#include <opencv2/opencv.hpp>
// #include <cv_bridge/cv_bridge.h>
// #include "sensor_msgs/image_encodings.h"

int main(int argc, char* argv[]) {
    std::string image_raw_topic = "/image_raw";
    std::string image_raw_compressed_topic = "/image_raw/compressed";

    std::vector<std::string> bag_topics;
    bag_topics.push_back(image_raw_topic);
    bag_topics.push_back(image_raw_compressed_topic);

    rosbag::Bag bag;
    bag.open("/home/alan/Desktop/alan_test.bag", rosbag::bagmode::Read);
    rosbag::View view(bag, rosbag::TopicQuery(bag_topics));
    // std::vector<std::string> types;
    // types.push_back(std::string("geometry_msgs/TransformStamped"));
    // rosbag::View view(bag, rosbag::TypeQuery(types));

    // sensor_msgs::CompressedImageConstPtr image_message = view.begin()->instantiate<sensor_msgs::CompressedImage>();
    
    for (const rosbag::MessageInstance& message_instance : view) {
        if (message_instance.getTopic() == image_raw_compressed_topic) {
            auto image_message = // sensor_msgs::CompressedImageConstPtr
                    message_instance.instantiate<sensor_msgs::CompressedImage>();
            long long timestamp = image_message->header.stamp.toNSec();
            // convert compressed image data to cv::Mat
            cv::Mat3b color_image =
                    cv::imdecode(cv::Mat(image_message->data), cv::IMREAD_COLOR); 
            // cv_bridge::CvImagePtr cv_ptr;
            // cv_ptr = cv_bridge::toCvCopy(image_message, sensor_msgs::image_encodings::BGR8);
            // cv::Mat color_image = cv_ptr->image;

            cv::imwrite("/home/alan/Desktop/image_raw/compressed/" + std::to_string(timestamp) + ".jpg", color_image);
            cv::waitKey(30);
        } else if (message_instance.getTopic() == image_raw_topic) {
            auto image_message =
                    message_instance.instantiate<sensor_msgs::CompressedImage>();
            long long timestamp = image_message->header.stamp.toNSec();
            cv::Mat3b color_image = cv::Mat(image_message->data);
            cv::imwrite("/home/alan/Desktop/image_raw/" + std::to_string(timestamp) + ".jpg", color_image);
            cv::waitKey(30);
        }
    }
}

解析pose

#include <geometry_msgs/PoseStamped.h>

else if (message_instance.getTopic() == config_.pose_topic()) {
      auto pose_message =
              message_instance.instantiate<geometry_msgs::PoseStamped>();
      ProcessPoseMessage(pose_message)

void ProcessPoseMessage(
  const geometry_msgs::PoseStamped::ConstPtr &pose_msg) {
  double timestamp = pose_msg->header.stamp.toSec();
  // double timestamp = pose_msg->header.stamp.sec + static_cast<double>(pose_msg->header.stamp.nsec) * 1e-9;
  // long long timestamp = pose_msg->header.stamp.toNSec();
  Eigen::Quaterniond orientation(pose_msg->pose.orientation.w,
                                  pose_msg->pose.orientation.x,
                                  pose_msg->pose.orientation.y,
                                  pose_msg->pose.orientation.z);
  Eigen::Vector3d position(pose_msg->pose.position.x,
                            pose_msg->pose.position.y,
                            pose_msg->pose.position.z);
}

3、topic重命名

main.cc:

#include <string>
#include <rosbag/bag.h>
#include <rosbag/view.h>

int main(int argc, char **argv) {
    std::string old_topic = "/old_topic_name";
    std::string new_topic = "/new_topic_name";
    std::string old_bag = "old.bag";
    std::string new_bag = "new.bag";
    rosbag::Bag bag_read;
    rosbag::Bag bag_write;
    bag_read.open(old_bag, rosbag::bagmode::Read);
    bag_write.open(new_bag, rosbag::bagmode::Write);
    rosbag::View view_all(bag_read);
    for(const rosbag::MessageInstance& m : view_all) {
        if (m.getTopic() == old_topic) {
            bag_write.write(new_topic, m.getTime(), m);
        } else {
            bag_write.write(m.getTopic(), m.getTime(), m);
        }
    }
    bag_read.close();
    bag_write.close();
    return 0;
}

CMakeLists.txt

cmake_minimum_required(VERSION 2.8.3)
project(topic_rename)

set(CMAKE_CXX_STANDARD 11)
set(EXECUTABLE_OUTPUT_PATH ${PROJECT_SOURCE_DIR}/bin)

find_package(catkin REQUIRED COMPONENTS
  roscpp
  rosbag
)

include_directories(
  ${catkin_INCLUDE_DIRS}
)

add_executable(topic_rename main.cc)
target_link_libraries(topic_rename
        ${catkin_LIBRARIES})

4、将一个bag中的topic内容加入到另一个bag中

(待改进：加入topic所有消息、写成模版、rosbag相似的命令行？)

#include <gflags/gflags.h>
#include <glog/logging.h>
#include <ros/ros.h>
#include <rosbag/bag.h>
#include <rosbag/view.h>
#include <std_msgs/String.h>

DEFINE_string(topic_name, "/foo/bar", "topic_name");
DEFINE_string(bag_path_in, "1.bag", "bag contain topic");
DEFINE_string(bag_path_out, "2.bag", "bag does not contain topic");

int main(int argc, char **argv)
{
  ros::init(argc, argv, "rosbag_add_topic");
  ros::NodeHandle nh;

  rosbag::Bag bag_in;
  bag_in.open(FLAGS_bag_path_in.c_str(), rosbag::bagmode::Read);
  rosbag::View view_in(bag_in, rosbag::TopicQuery(FLAGS_topic_name));
  if (view_in.size() == 0) {
    LOG(FATAL) << FLAGS_bag_path_in << " does not contain topic: " << FLAGS_topic_name;
  }
  rosbag::Bag bag_out;
  bag_out.open(FLAGS_bag_path_out.c_str(), rosbag::bagmode::Append);

  std_msgs::String string_message = *(view_in.begin()->instantiate<std_msgs::String>());
  // std_msgs::String string_message;
  // string_message.data = std::string("foo");
  bag_out.write(FLAGS_topic_name, view_in.getBeginTime(), string_message);

  bag_in.close();
  bag_out.close();
  return 0;
}

三、Python

1、topic重命名

#!/usr/bin/env python3

import os
import argparse
import logging
import rosbag

def run():
    logging.basicConfig(format='%(levelname)s: %(message)s', level=logging.INFO)

    parser = argparse.ArgumentParser(description='......')
    parser.add_argument('-old_topic', default='/old_topic_name')
    parser.add_argument('-new_topic', default='/new_topic_name')
    parser.add_argument('-b', dest='bag', required=True)
    parser.add_argument('-o', dest='output_bag', default='')
    args = parser.parse_args()

    output_bag = args.output_bag
    if not output_bag or output_bag == args.bag:
        strs=os.path.splitext(args.bag)
        output_bag = strs[0] + '_' + strs[-1]

    bag_read = rosbag.Bag(args.bag)
    bag_write = rosbag.Bag(output_bag, mode='w')
    for topic, msg, t in bag_read.read_messages(): # 指定topic bag_read.read_messages(topics=['/foo', '/bar'])
        # time = t.to_sec()
        # x = msg.pose.position.x
        if topic == args.old_topic:
            bag_write.write(args.new_topic, msg, t)
        else:
            bag_write.write(topic, msg, t)
    bag_read.close()
    bag_write.close()

    logging.info("Done!\ntopic rename: %s --> %s\noutput bag: %s" % (args.old_topic, args.new_topic, output_bag))

if __name__ == '__main__':
    run()

参考：

http://wiki.ros.org/rosbag/Commandline

http://wiki.ros.org/cv_bridge/Tutorials/UsingCvBridgeToConvertBetweenROSImagesAndOpenCVImages

在ROS中 opencv 发布和接收图像消息

rosbag-数据记录工具

rosbag::Bag bag;
bag.open("/tmp/test.bag", rosbag::bagmode::Write);

void PoseCallback(const ros::MessageEvent<const geometry_msgs::PoseStamped> &event) {
  bag.write("/pose_topic", event.getReceiptTime(), event.getMessage());
}

int main(int argc, char **argv) {
  ros::init(argc, argv, "record_node");
  ros::NodeHandle nh("~");
  ros::Subscriber pose_sub_ = nh.subscribe("/pose_topic", 1, &PoseCallback);
  ros::spin();
}

Alan Lan

关注

4
点赞
踩
28

收藏

觉得还不错? 一键收藏
1
评论
rosbag相关

一、录制图像bag：#include <ros/ros.h>#include <rosbag/bag.h>#include <opencv2/highgui/highgui.hpp>#include "cv_bridge/cv_bridge.h" // opencv图像格式转化为ROS图像格式int main(int argc, char **...
复制链接

扫一扫

专栏目录