kitti数据可视化

半路程序员2024-02-28 8:13

数据下载

The KITTI Vision Benchmark Suite

这里以 2011_09_26_drive_0005 (0.6 GB)数据为参考，下载[synced+rectified data] [calibration] 数据。

下载完毕之后解压，然后将calibration文件解压后的结果放在如下目录下，

下载kitti2bag包

python 复制代码 
pip install kitti2bag

如果后期出现如下错误：

ImportError: dynamic module does not define module export function (PyInit__tf2)

则执行如下重新安装

python 复制代码 
pip2 install kitti2bag

进入到2011_09_26文件夹的同级目录中，执行如下代码进行转换：

bash 复制代码 
kitti2bag -t 2011_09_26 -r 0005 raw_synced

运行截图如下：

运行结束之后，会在当前目录下生成一个bag文件。

启动ROS，并进行播放

bash 复制代码 
roscore

进入到刚刚生成bag文件的目录下，执行如下代码：

bash 复制代码 
rqt_bag kitti_2011_09_26_drive_0005_synced.bag

生成如下截图：

选择其中一个图像话题，右键，选择view -》选择image，然后点击play进行播放。

发布图片

创建工作空间，并创建功能包

bash 复制代码 
~/catkin_ws/src$ catkin_create_pkg kitti_tutorial rospy

编写kitti.py代码

python 复制代码 
#!/usr/bin/env python

import cv2
import os

import rospy
from sensor_msgs.msg import Image
from cv_bridge import CvBridge

DATA_PATH = "/home/d501/kitti_data/2011_09_26/2011_09_26_drive_0005_sync/2011_09_26/2011_09_26_drive_0005_sync"

if __name__ == "__main__":
    frame = 0
    rospy.init_node("kitti_node", anonymous=True)
    cam_pub = rospy.Publisher('kitti_cam', Image, queue_size=10)
    bridge = CvBridge()

    rate = rospy.Rate(10)
    while not rospy.is_shutdown():
        img = cv2.imread(os.path.join(DATA_PATH, 'image_02/data/%010d.png'%frame))
        cam_pub.publish(bridge.cv2_to_imgmsg(img, 'bgr8'))
        rospy.loginfo("camera image published")
        rate.sleep()
        frame += 1
        frame %= 154

添加可执行权限

bash 复制代码 
sudo chmod +x kitti.py

编译工作空间

bash 复制代码 
~/catkin_ws$ catkin_make
~/catkin_ws$ source devel/setup.bash

在使用pycharm编写python程序时，

问题：在pycharm中使用ros时候，遇到"No module named rospy"

pycharm中无法加载rospy_pycharm no module named rospy-CSDN博客

将解释器更换为opt下的python2.7

启动rviz

bash 复制代码 
rviz

启动ros节点

bash 复制代码 
~/catkin_ws$ rosrun kitti_tutorial kitti.py

添加topic

如图所示：

发布点云数据

修改kitti.py代码

python 复制代码 
#!/usr/bin/env python
import numpy as np

import cv2
import os
import numpy as np
import rospy
from std_msgs.msg import Header
from sensor_msgs.msg import Image, PointCloud2
import sensor_msgs.point_cloud2 as pcl2
from cv_bridge import CvBridge

DATA_PATH = "/home/d501/kitti_data/2011_09_26/2011_09_26_drive_0005_sync/2011_09_26/2011_09_26_drive_0005_sync"

if __name__ == "__main__":
    frame = 0
    rospy.init_node("kitti_node", anonymous=True)
    cam_pub = rospy.Publisher('kitti_cam', Image, queue_size=10)
    pcl_pub = rospy.Publisher('kitti_point_cloud', PointCloud2, queue_size=10)
    bridge = CvBridge()

    rate = rospy.Rate(10)
    while not rospy.is_shutdown():
        img = cv2.imread(os.path.join(DATA_PATH, 'image_02/data/%010d.png'%frame))
        point_cloud = np.fromfile(os.path.join(DATA_PATH, 'velodyne_points/data/%010d.bin'%frame), dtype=np.float32).reshape(-1, 4)
        cam_pub.publish(bridge.cv2_to_imgmsg(img, 'bgr8'))

        header = Header()
        header.stamp = rospy.Time.now()
        header.frame_id = 'map'
        pcl_pub.publish(pcl2.create_cloud_xyz32(header, point_cloud[:,:3]))
        rospy.loginfo("published")
        rate.sleep()
        frame += 1
        frame %= 154

重新添加topic之后，如下：

调整size大小

将rviz存档，以后开启它就不用重新添加topic

点击File->Save Config As->选择包下的某个路径，保存为xxxx.rviz文件，下次启动时，使用rviz -d xxxx.rviz即可

添加照相机视野

重构代码data_utils.py

python 复制代码 
#!/usr/bin/env python

import cv2
import numpy as np

def read_camera(path):
    return cv2.imread(path)

def read_point_cloud(path):
    return np.fromfile(path, dtype=np.float32).reshape(-1, 4)

publish_utils.py

python 复制代码 
#!/usr/bin/env python

import rospy
from geometry_msgs.msg import Point
from std_msgs.msg import Header
from sensor_msgs.msg import Image, PointCloud2
import sensor_msgs.point_cloud2 as pcl2
from cv_bridge import CvBridge
from visualization_msgs.msg import Marker

FRAME_ID = 'map'

def publish_camera(cam_pub, bridge, image):
    cam_pub.publish(bridge.cv2_to_imgmsg(image, 'bgr8'))

def publish_point_cloud(pcl_pub, point_cloud):
    header = Header()
    header.stamp = rospy.Time.now()
    header.frame_id = FRAME_ID
    pcl_pub.publish(pcl2.create_cloud_xyz32(header, point_cloud[:, :3]))

def publish_ego_car(ego_car_pub):
    """
    publish left and right 45 degree FOV lines and ego car model mesh
    :param ego_car_pub:
    :return:
    """
    marker = Marker()

    marker.header.frame_id = FRAME_ID
    marker.header.stamp = rospy.Time.now()

    marker.id = 0
    marker.action = Marker.ADD
    marker.lifetime = rospy.Duration()
    marker.type = Marker.LINE_STRIP

    marker.color.r = 0.0
    marker.color.g = 1.0
    marker.color.b = 0.0
    marker.color.a = 1.0
    marker.scale.x = 0.2

    marker.points = []
    marker.points.append(Point(10, -10, 0))
    marker.points.append(Point(0, 0, 0))
    marker.points.append(Point(10, 10, 0))

    ego_car_pub.publish(marker)

kitti.py

python 复制代码 
#!/usr/bin/env python
import numpy as np

import rospy

from data_utils import *
from publish_utils import *
import os

DATA_PATH = "/home/d501/kitti_data/2011_09_26/2011_09_26_drive_0005_sync/2011_09_26/2011_09_26_drive_0005_sync"

if __name__ == "__main__":
    frame = 0
    rospy.init_node("kitti_node", anonymous=True)
    cam_pub = rospy.Publisher('kitti_cam', Image, queue_size=10)
    pcl_pub = rospy.Publisher('kitti_point_cloud', PointCloud2, queue_size=10)
    ego_pub = rospy.Publisher('kitti_ego_car', Marker, queue_size=10)

    bridge = CvBridge()

    rate = rospy.Rate(10)
    while not rospy.is_shutdown():
        image = read_camera(os.path.join(DATA_PATH, 'image_02/data/%010d.png'%frame))
        point_cloud = read_point_cloud(os.path.join(DATA_PATH, 'velodyne_points/data/%010d.bin'%frame))

        publish_camera(cam_pub, bridge, image)
        publish_point_cloud(pcl_pub, point_cloud)
        publish_ego_car(ego_pub)

        rospy.loginfo("published")
        rate.sleep()
        frame += 1
        frame %= 154

画出汽车模型

下载汽车模型Low-Poly Car Free 3D Model - .3ds .obj .dae .blend .fbx .mtl - Free3D

解压之后，将Car.dae文件放入到包下的某个文件夹中，这里直接使用原生解压的文件夹目录。

在publish_utils.py中添加如下代码

python 复制代码 
def publish_car_model(car_model_pub):
    mech_marker = Marker()
    mech_marker.header.frame_id = "map"
    mech_marker.header.stamp = rospy.Time.now()
    # id必须不一样，且只能是整数
    mech_marker.id = -1
    mech_marker.type = Marker.MESH_RESOURCE
    mech_marker.action = Marker.ADD
    mech_marker.lifetime = rospy.Duration()
    mech_marker.mesh_resource = "package://kitti_tutorials/am7f3psa0agw-Car-Model/Car-Model/Car.dae"  # .dae模型文件的相对地址

    # 位置主要看官方提供的位置图
    # 因为自车的velodyne激光雷达相对于map的位置是（0，0，0），看设备安装图上velodyne的位置是（0，0，1.73），显而易见，在rviz中自车模型位置应该是（0，0，-1.73）
    mech_marker.pose.position.x = 0.0
    mech_marker.pose.position.y = 0.0
    mech_marker.pose.position.z = -1.73

    # 这边和原作者的代码不一样，因为tf结构发生了一些改变，四元数需要自己根据模型做响应的调整，笔者这边是调整好的模型四元数
    q = transformations.quaternion_from_euler(np.pi, np.pi, -np.pi / 2.0)
    mech_marker.pose.orientation.x = q[0]
    mech_marker.pose.orientation.y = q[1]
    mech_marker.pose.orientation.z = q[2]
    mech_marker.pose.orientation.w = q[3]

    mech_marker.color.r = 1.0
    mech_marker.color.g = 1.0
    mech_marker.color.b = 1.0
    mech_marker.color.a = 1.0  # 透明度

    # 设置车辆大小
    mech_marker.scale.x = 1.0
    mech_marker.scale.y = 1.0
    mech_marker.scale.z = 1.0

    car_model_pub.publish(mech_marker)

修改kitti.py代码

python 复制代码 
#!/usr/bin/env python
import numpy as np

import rospy

from data_utils import *
from publish_utils import *
import os

DATA_PATH = "/home/d501/kitti_data/2011_09_26/2011_09_26_drive_0005_sync/2011_09_26/2011_09_26_drive_0005_sync"

if __name__ == "__main__":
    frame = 0
    rospy.init_node("kitti_node", anonymous=True)
    cam_pub = rospy.Publisher('kitti_cam', Image, queue_size=10)
    pcl_pub = rospy.Publisher('kitti_point_cloud', PointCloud2, queue_size=10)
    ego_pub = rospy.Publisher('kitti_ego_car', Marker, queue_size=10)
    model_pub = rospy.Publisher('kitti_car_model', Marker, queue_size=10)

    bridge = CvBridge()

    rate = rospy.Rate(10)
    while not rospy.is_shutdown():
        image = read_camera(os.path.join(DATA_PATH, 'image_02/data/%010d.png'%frame))
        point_cloud = read_point_cloud(os.path.join(DATA_PATH, 'velodyne_points/data/%010d.bin'%frame))

        publish_camera(cam_pub, bridge, image)
        publish_point_cloud(pcl_pub, point_cloud)
        publish_ego_car(ego_pub)
        publish_car_model(model_pub)

        rospy.loginfo("published")
        rate.sleep()
        frame += 1
        frame %= 154

重新运行之后，出现错误：

[ERROR] [1679018952.335224247]: Could not load resource [package://kitti_tutorials/meshes/Car.dae]: Unable to open file "package://kitti_tutorials/meshes/Car.dae".

[rospack] Error: package 'kitti_tutorial' not found

参考：rviz进行kitti数据集可视化时加载小车模型报错_rviz could not load resourceunable to open file-CSDN博客

使用MarkerArray进行发布

修改publish_utils.py

python 复制代码 
def publish_ego_car(ego_car_pub):
    """
    publish left and right 45 degree FOV lines and ego car model mesh
    :param ego_car_pub:
    :return:
    """
    marker_array = MarkerArray()
    marker = Marker()

    marker.header.frame_id = FRAME_ID
    marker.header.stamp = rospy.Time.now()

    marker.id = 0
    marker.action = Marker.ADD
    marker.lifetime = rospy.Duration()
    marker.type = Marker.LINE_STRIP

    marker.color.r = 0.0
    marker.color.g = 1.0
    marker.color.b = 0.0
    marker.color.a = 1.0
    marker.scale.x = 0.2

    marker.points = []
    marker.points.append(Point(10, -10, 0))
    marker.points.append(Point(0, 0, 0))
    marker.points.append(Point(10, 10, 0))

    marker_array.markers.append(marker)

    mech_marker = Marker()
    mech_marker.header.frame_id = "map"
    mech_marker.header.stamp = rospy.Time.now()
    mech_marker.id = -1
    mech_marker.type = Marker.MESH_RESOURCE
    mech_marker.action = Marker.ADD
    mech_marker.lifetime = rospy.Duration()
    mech_marker.mesh_resource = "package://kitti_tutorial/meshes/Car.dae"

    mech_marker.pose.position.x = 0.0
    mech_marker.pose.position.y = 0.0
    mech_marker.pose.position.z = -1.73

    q = transformations.quaternion_from_euler(np.pi, np.pi, -np.pi / 2.0)
    mech_marker.pose.orientation.x = q[0]
    mech_marker.pose.orientation.y = q[1]
    mech_marker.pose.orientation.z = q[2]
    mech_marker.pose.orientation.w = q[3]

    mech_marker.color.r = 1.0
    mech_marker.color.g = 1.0
    mech_marker.color.b = 1.0
    mech_marker.color.a = 1.0

    mech_marker.scale.x = 1.0
    mech_marker.scale.y = 1.0
    mech_marker.scale.z = 1.0

    marker_array.markers.append(mech_marker)

    ego_car_pub.publish(marker_array)

kitti.py代码

python 复制代码 
#!/usr/bin/env python
import numpy as np

import rospy

from data_utils import *
from publish_utils import *
import os

DATA_PATH = "/home/d501/kitti_data/2011_09_26/2011_09_26_drive_0005_sync/2011_09_26/2011_09_26_drive_0005_sync"

if __name__ == "__main__":
    frame = 0
    rospy.init_node("kitti_node", anonymous=True)
    cam_pub = rospy.Publisher('kitti_cam', Image, queue_size=10)
    pcl_pub = rospy.Publisher('kitti_point_cloud', PointCloud2, queue_size=10)
    ego_pub = rospy.Publisher('kitti_ego_car', MarkerArray, queue_size=10)

    bridge = CvBridge()

    rate = rospy.Rate(10)
    while not rospy.is_shutdown():
        image = read_camera(os.path.join(DATA_PATH, 'image_02/data/%010d.png'%frame))
        point_cloud = read_point_cloud(os.path.join(DATA_PATH, 'velodyne_points/data/%010d.bin'%frame))

        publish_camera(cam_pub, bridge, image)
        publish_point_cloud(pcl_pub, point_cloud)
        publish_ego_car(ego_pub)

        rospy.loginfo("published")
        rate.sleep()
        frame += 1
        frame %= 154

发布IMU数据

data_utils.py添加

python 复制代码 
IMU_NAME = ["lat", "lon", "alt", "roll", "pitch", "yaw", "vn", "ve", "vf", "vl", "vu", "ax", "ay", "az", "af", "al", "au", "wx", "wy", "wz", "wf", "wl", "wu", "posacc", "velacc", "navstat", "numsats", "posmode", "velmode", "orimode"]

def read_imu(path):
    imu_data = pd.read_csv(path, header=None, sep = " ")
    imu_data.columns = IMU_NAME
    return imu_data

publish_utils.py

python 复制代码 
def publish_imu(imu_pub, imu_data):
    imu = Imu()

    imu.header.frame_id = "map"
    imu.header.stamp = rospy.Time.now()

    q = transformations.quaternion_from_euler(float(imu_data.roll), float(imu_data.pitch), float(imu_data.yaw))
    imu.orientation.x = q[0]
    imu.orientation.y = q[1]
    imu.orientation.z = q[2]
    imu.orientation.w = q[3]

    imu.linear_acceleration.x = imu_data.af
    imu.linear_acceleration.y = imu_data.al
    imu.linear_acceleration.z = imu_data.au

    imu.angular_velocity.x = imu_data.wx
    imu.angular_velocity.y = imu_data.wy
    imu.angular_velocity.z = imu_data.wz

    imu_pub.publish(imu)

kitti.py

python 复制代码 
#!/usr/bin/env python
import numpy as np

import rospy

from data_utils import *
from publish_utils import *
import os

DATA_PATH = "/home/d501/kitti_data/2011_09_26/2011_09_26_drive_0005_sync/2011_09_26/2011_09_26_drive_0005_sync"

if __name__ == "__main__":
    frame = 0
    rospy.init_node("kitti_node", anonymous=True)
    cam_pub = rospy.Publisher('kitti_cam', Image, queue_size=10)
    pcl_pub = rospy.Publisher('kitti_point_cloud', PointCloud2, queue_size=10)
    ego_pub = rospy.Publisher('kitti_ego_car', MarkerArray, queue_size=10)
    imu_pub = rospy.Publisher('kitti_imu', Imu, queue_size=10)

    bridge = CvBridge()

    rate = rospy.Rate(10)
    while not rospy.is_shutdown():
        image = read_camera(os.path.join(DATA_PATH, 'image_02/data/%010d.png'%frame))
        point_cloud = read_point_cloud(os.path.join(DATA_PATH, 'velodyne_points/data/%010d.bin'%frame))
        imu_data = read_imu(os.path.join(DATA_PATH, 'oxts/data/%010d.txt'%frame))

        publish_camera(cam_pub, bridge, image)
        publish_point_cloud(pcl_pub, point_cloud)
        publish_ego_car(ego_pub)
        publish_imu(imu_pub, imu_data)

        rospy.loginfo("published")
        rate.sleep()
        frame += 1
        frame %= 154

订阅话题显示如下：

发布GPS数据

publish_utils.py添加代码

python 复制代码 
def publish_gps(gps_pub, gps_data):
    gps = NavSatFix()

    gps.header.frame_id = "map"
    gps.header.stamp = rospy.Time.now()

    gps.latitude = gps_data.lat
    gps.longitude = gps_data.lon
    gps.altitude = gps_data.alt

    gps_pub.publish(gps)

kitti.py代码

python 复制代码 
#!/usr/bin/env python
import numpy as np

import rospy

from data_utils import *
from publish_utils import *
import os

DATA_PATH = "/home/d501/kitti_data/2011_09_26/2011_09_26_drive_0005_sync/2011_09_26/2011_09_26_drive_0005_sync"

if __name__ == "__main__":
    frame = 0
    rospy.init_node("kitti_node", anonymous=True)
    cam_pub = rospy.Publisher('kitti_cam', Image, queue_size=10)
    pcl_pub = rospy.Publisher('kitti_point_cloud', PointCloud2, queue_size=10)
    ego_pub = rospy.Publisher('kitti_ego_car', MarkerArray, queue_size=10)
    imu_pub = rospy.Publisher('kitti_imu', Imu, queue_size=10)
    gps_pub = rospy.Publisher('kitti_gps', NavSatFix, queue_size=10)

    bridge = CvBridge()

    rate = rospy.Rate(10)
    while not rospy.is_shutdown():
        image = read_camera(os.path.join(DATA_PATH, 'image_02/data/%010d.png'%frame))
        point_cloud = read_point_cloud(os.path.join(DATA_PATH, 'velodyne_points/data/%010d.bin'%frame))
        imu_data = read_imu(os.path.join(DATA_PATH, 'oxts/data/%010d.txt'%frame))

        publish_camera(cam_pub, bridge, image)
        publish_point_cloud(pcl_pub, point_cloud)
        publish_ego_car(ego_pub)
        publish_imu(imu_pub, imu_data)
        publish_gps(gps_pub, imu_data)

        rospy.loginfo("published")
        rate.sleep()
        frame += 1
        frame %= 154

画出2D检测框

参考ROS1结合自动驾驶数据集Kitti开发教程(七)下载图像标注资料并读取显示_kitti 视觉slam00/data.txt-CSDN博客

ROS1结合自动驾驶数据集Kitti开发教程(八)画出2D检测框_ros2如何在image上绘制2d包围框-CSDN博客

代码：

kitti.py

python 复制代码 
#!/usr/bin/env python
import numpy as np

import rospy

from data_utils import *
from publish_utils import *
import os

DATA_PATH = "/home/d501/kitti_data/2011_09_26/2011_09_26_drive_0005_sync/2011_09_26/2011_09_26_drive_0005_sync"

if __name__ == "__main__":
    frame = 0
    rospy.init_node("kitti_node", anonymous=True)
    cam_pub = rospy.Publisher('kitti_cam', Image, queue_size=10)
    pcl_pub = rospy.Publisher('kitti_point_cloud', PointCloud2, queue_size=10)
    ego_pub = rospy.Publisher('kitti_ego_car', MarkerArray, queue_size=10)
    imu_pub = rospy.Publisher('kitti_imu', Imu, queue_size=10)
    gps_pub = rospy.Publisher('kitti_gps', NavSatFix, queue_size=10)

    bridge = CvBridge()

    df_tracking = read_labelfile('/home/d501//kitti_data/2011_09_26/data_tracking_label_2/training/label_02/0000.txt')

    rate = rospy.Rate(10)
    while not rospy.is_shutdown():
        boxes = np.array(df_tracking[df_tracking.frame==frame][['bbox_left','bbox_top','bbox_right','bbox_bottom']])
        types = np.array(df_tracking[df_tracking.frame==frame]['type'])

        image = read_camera(os.path.join(DATA_PATH, 'image_02/data/%010d.png'%frame))
        point_cloud = read_point_cloud(os.path.join(DATA_PATH, 'velodyne_points/data/%010d.bin'%frame))
        imu_data = read_imu(os.path.join(DATA_PATH, 'oxts/data/%010d.txt'%frame))

        publish_camera(cam_pub, bridge, image, boxes, types)
        publish_point_cloud(pcl_pub, point_cloud)
        publish_ego_car(ego_pub)
        publish_imu(imu_pub, imu_data)
        publish_gps(gps_pub, imu_data)

        rospy.loginfo("published")
        rate.sleep()
        frame += 1
        frame %= 154

data_utils.py

python 复制代码 
#!/usr/bin/env python

import cv2
import numpy as np
import pandas as pd

IMU_NAME = ["lat", "lon", "alt", "roll", "pitch", "yaw", "vn", "ve", "vf", "vl", "vu", "ax", "ay", "az", "af", "al", "au", "wx", "wy", "wz", "wf", "wl", "wu", "posacc", "velacc", "navstat", "numsats", "posmode", "velmode", "orimode"]

LABEL_NAME = ["frame", "track id", "type", "truncated", "occluded", "alpha", "bbox_left", "bbox_top", "bbox_right", "bbox_bottom", "dimensions_height", "dimensions_width", "dimensions_length", "location_x", "location_y", "location_z", "rotation_y"]

def read_camera(path):
    return cv2.imread(path)

def read_point_cloud(path):
    return np.fromfile(path, dtype=np.float32).reshape(-1, 4)

def read_imu(path):
    imu_data = pd.read_csv(path, header=None, sep = " ")
    imu_data.columns = IMU_NAME
    return imu_data


def read_labelfile(path):
    data = pd.read_csv(path, header=None, sep=" ")
    data.columns = LABEL_NAME
    data.loc[data.type.isin(['Van', 'Car', 'Truck']), 'type'] = 'Car'
    data = data[data.type.isin(['Car', 'Cyclist', 'Pedestrian'])]
    return data

publish_utils.py

python 复制代码 
#!/usr/bin/env python
import cv2

import rospy
from geometry_msgs.msg import Point
from std_msgs.msg import Header
from sensor_msgs.msg import Image, PointCloud2, Imu, NavSatFix
import sensor_msgs.point_cloud2 as pcl2
from cv_bridge import CvBridge
from tf import transformations
from visualization_msgs.msg import Marker, MarkerArray
import numpy as np
import pandas as pd
import os
FRAME_ID = 'map'

DETECTION_COLOR_DICT = {'Car': (255, 255, 0), 'Pedestrian': (0, 226, 255), 'Cyclist': (141, 40, 255)}

def publish_camera(cam_pub, bridge, image, boxes, types):
    for type, box in zip(types, boxes):
        top_left = int(box[0]), int(box[1])
        bottom_right = int(box[2]), int(box[3])
        cv2.rectangle(image, top_left, bottom_right, DETECTION_COLOR_DICT[type], 2)
    cam_pub.publish(bridge.cv2_to_imgmsg(image, 'bgr8'))

def publish_point_cloud(pcl_pub, point_cloud):
    header = Header()
    header.stamp = rospy.Time.now()
    header.frame_id = FRAME_ID
    pcl_pub.publish(pcl2.create_cloud_xyz32(header, point_cloud[:, :3]))

def publish_ego_car(ego_car_pub):
    """
    publish left and right 45 degree FOV lines and ego car model mesh
    :param ego_car_pub:
    :return:
    """
    marker_array = MarkerArray()
    marker = Marker()

    marker.header.frame_id = FRAME_ID
    marker.header.stamp = rospy.Time.now()

    marker.id = 0
    marker.action = Marker.ADD
    marker.lifetime = rospy.Duration()
    marker.type = Marker.LINE_STRIP

    marker.color.r = 0.0
    marker.color.g = 1.0
    marker.color.b = 0.0
    marker.color.a = 1.0
    marker.scale.x = 0.2

    marker.points = []
    marker.points.append(Point(10, -10, 0))
    marker.points.append(Point(0, 0, 0))
    marker.points.append(Point(10, 10, 0))

    marker_array.markers.append(marker)

    mech_marker = Marker()
    mech_marker.header.frame_id = "map"
    mech_marker.header.stamp = rospy.Time.now()
    mech_marker.id = -1
    mech_marker.type = Marker.MESH_RESOURCE
    mech_marker.action = Marker.ADD
    mech_marker.lifetime = rospy.Duration()
    mech_marker.mesh_resource = "package://kitti_tutorial/meshes/Car.dae"

    mech_marker.pose.position.x = 0.0
    mech_marker.pose.position.y = 0.0
    mech_marker.pose.position.z = -1.73

    q = transformations.quaternion_from_euler(np.pi, np.pi, -np.pi / 2.0)
    mech_marker.pose.orientation.x = q[0]
    mech_marker.pose.orientation.y = q[1]
    mech_marker.pose.orientation.z = q[2]
    mech_marker.pose.orientation.w = q[3]

    mech_marker.color.r = 1.0
    mech_marker.color.g = 1.0
    mech_marker.color.b = 1.0
    mech_marker.color.a = 1.0

    mech_marker.scale.x = 1.0
    mech_marker.scale.y = 1.0
    mech_marker.scale.z = 1.0

    marker_array.markers.append(mech_marker)

    ego_car_pub.publish(marker_array)


def publish_imu(imu_pub, imu_data):
    imu = Imu()

    imu.header.frame_id = "map"
    imu.header.stamp = rospy.Time.now()

    q = transformations.quaternion_from_euler(float(imu_data.roll), float(imu_data.pitch), float(imu_data.yaw))
    imu.orientation.x = q[0]
    imu.orientation.y = q[1]
    imu.orientation.z = q[2]
    imu.orientation.w = q[3]

    imu.linear_acceleration.x = imu_data.af
    imu.linear_acceleration.y = imu_data.al
    imu.linear_acceleration.z = imu_data.au

    imu.angular_velocity.x = imu_data.wx
    imu.angular_velocity.y = imu_data.wy
    imu.angular_velocity.z = imu_data.wz

    imu_pub.publish(imu)

def publish_gps(gps_pub, gps_data):
    gps = NavSatFix()

    gps.header.frame_id = "map"
    gps.header.stamp = rospy.Time.now()

    gps.latitude = gps_data.lat
    gps.longitude = gps_data.lon
    gps.altitude = gps_data.alt

    gps_pub.publish(gps)

结果如下：

画出3D检测框

参考：ROS1结合自动驾驶数据集Kitti开发教程(九)画出3D检测框[1]_ros 画3d框-CSDN博客

下面贴出代码：

publish_utils.py

python 复制代码 
#!/usr/bin/env python
import cv2

import rospy
from geometry_msgs.msg import Point
from std_msgs.msg import Header
from sensor_msgs.msg import Image, PointCloud2, Imu, NavSatFix
import sensor_msgs.point_cloud2 as pcl2
from cv_bridge import CvBridge
from tf import transformations
from visualization_msgs.msg import Marker, MarkerArray
import numpy as np
import pandas as pd
import os
FRAME_ID = 'map'

DETECTION_COLOR_DICT = {'Car': (255, 255, 0), 'Pedestrian': (0, 226, 255), 'Cyclist': (141, 40, 255)}
LIFETIME = 0.1
LINES = [[0, 1], [1, 2], [2, 3], [3, 0]]  # Lower plane parallel to Z=0 plane
LINES += [[4, 5], [5, 6], [6, 7], [7, 4]]  # Upper plane parallel to Z=0 plane
LINES += [[0, 4], [1, 5], [2, 6], [3, 7]]  # Connections between upper and lower planes
LINES += [[4,1], [5,0]]

def publish_camera(cam_pub, bridge, image, boxes, types):
    for type, box in zip(types, boxes):
        top_left = int(box[0]), int(box[1])
        bottom_right = int(box[2]), int(box[3])
        cv2.rectangle(image, top_left, bottom_right, DETECTION_COLOR_DICT[type], 2)
    cam_pub.publish(bridge.cv2_to_imgmsg(image, 'bgr8'))

def publish_point_cloud(pcl_pub, point_cloud):
    header = Header()
    header.stamp = rospy.Time.now()
    header.frame_id = FRAME_ID
    pcl_pub.publish(pcl2.create_cloud_xyz32(header, point_cloud[:, :3]))

def publish_ego_car(ego_car_pub):
    """
    publish left and right 45 degree FOV lines and ego car model mesh
    :param ego_car_pub:
    :return:
    """
    marker_array = MarkerArray()
    marker = Marker()

    marker.header.frame_id = FRAME_ID
    marker.header.stamp = rospy.Time.now()

    marker.id = 0
    marker.action = Marker.ADD
    marker.lifetime = rospy.Duration()
    marker.type = Marker.LINE_STRIP

    marker.color.r = 0.0
    marker.color.g = 1.0
    marker.color.b = 0.0
    marker.color.a = 1.0
    marker.scale.x = 0.2

    marker.points = []
    marker.points.append(Point(10, -10, 0))
    marker.points.append(Point(0, 0, 0))
    marker.points.append(Point(10, 10, 0))

    marker_array.markers.append(marker)

    mech_marker = Marker()
    mech_marker.header.frame_id = "map"
    mech_marker.header.stamp = rospy.Time.now()
    mech_marker.id = -1
    mech_marker.type = Marker.MESH_RESOURCE
    mech_marker.action = Marker.ADD
    mech_marker.lifetime = rospy.Duration()
    mech_marker.mesh_resource = "package://kitti_tutorial/meshes/Car.dae"

    mech_marker.pose.position.x = 0.0
    mech_marker.pose.position.y = 0.0
    mech_marker.pose.position.z = -1.73

    q = transformations.quaternion_from_euler(np.pi, np.pi, -np.pi / 2.0)
    mech_marker.pose.orientation.x = q[0]
    mech_marker.pose.orientation.y = q[1]
    mech_marker.pose.orientation.z = q[2]
    mech_marker.pose.orientation.w = q[3]

    mech_marker.color.r = 1.0
    mech_marker.color.g = 1.0
    mech_marker.color.b = 1.0
    mech_marker.color.a = 1.0

    mech_marker.scale.x = 1.0
    mech_marker.scale.y = 1.0
    mech_marker.scale.z = 1.0

    marker_array.markers.append(mech_marker)

    ego_car_pub.publish(marker_array)


def publish_imu(imu_pub, imu_data):
    imu = Imu()

    imu.header.frame_id = "map"
    imu.header.stamp = rospy.Time.now()

    q = transformations.quaternion_from_euler(float(imu_data.roll), float(imu_data.pitch), float(imu_data.yaw))
    imu.orientation.x = q[0]
    imu.orientation.y = q[1]
    imu.orientation.z = q[2]
    imu.orientation.w = q[3]

    imu.linear_acceleration.x = imu_data.af
    imu.linear_acceleration.y = imu_data.al
    imu.linear_acceleration.z = imu_data.au

    imu.angular_velocity.x = imu_data.wx
    imu.angular_velocity.y = imu_data.wy
    imu.angular_velocity.z = imu_data.wz

    imu_pub.publish(imu)

def publish_gps(gps_pub, gps_data):
    gps = NavSatFix()

    gps.header.frame_id = "map"
    gps.header.stamp = rospy.Time.now()

    gps.latitude = gps_data.lat
    gps.longitude = gps_data.lon
    gps.altitude = gps_data.alt

    gps_pub.publish(gps)


def publish_3dbox(box3d_pub, corners_3d_velos, types):
    marker_array = MarkerArray()
    for i, corners_3d_velo in enumerate(corners_3d_velos):
        marker = Marker()
        marker.header.frame_id = FRAME_ID
        marker.header.stamp = rospy.Time.now()

        marker.id = i
        marker.action = Marker.ADD
        marker.lifetime = rospy.Duration(LIFETIME)
        marker.type = Marker.LINE_LIST

        b, g, r = DETECTION_COLOR_DICT[types[i]]
        marker.color.r = r/255.0
        marker.color.g = g/255.0
        marker.color.b = b/255.0

        marker.color.a = 1.0
        marker.scale.x = 0.1

        marker.points = []
        for l in LINES:
            p1 = corners_3d_velo[l[0]]
            marker.points.append(Point(p1[0], p1[1], p1[2]))
            p2 = corners_3d_velo[l[1]]
            marker.points.append(Point(p2[0], p2[1], p2[2]))
        marker_array.markers.append(marker)

    box3d_pub.publish(marker_array)

下载calibration转换程序：kitti_util.py

kitti.py

python 复制代码 
#!/usr/bin/env python
#  -*-coding:utf8 -*-
import numpy as np

import rospy

from data_utils import *
from publish_utils import *
from kitti_util import *
import os

DATA_PATH = "/home/d501/kitti_data/2011_09_26/2011_09_26_drive_0005_sync/2011_09_26/2011_09_26_drive_0005_sync"

def compute_3d_box_cam2(h, w, l, x, y, z, yaw):
    # 计算旋转矩阵
    R = np.array([[np.cos(yaw), 0, np.sin(yaw)], [0, 1, 0], [-np.sin(yaw), 0, np.cos(yaw)]])
    # 8个顶点的xyz
    x_corners = [l/2,l/2,-l/2,-l/2,l/2,l/2,-l/2,-l/2]
    y_corners = [0,0,0,0,-h,-h,-h,-h]
    z_corners = [w/2,-w/2,-w/2,w/2,w/2,-w/2,-w/2,w/2]
    # 旋转矩阵点乘(3，8)顶点矩阵
    corners_3d_cam2 = np.dot(R, np.vstack([x_corners,y_corners,z_corners]))
    # 加上location中心点，得出8个顶点旋转后的坐标
    corners_3d_cam2 += np.vstack([x,y,z])
    return corners_3d_cam2


if __name__ == "__main__":
    frame = 0
    rospy.init_node("kitti_node", anonymous=True)
    cam_pub = rospy.Publisher('kitti_cam', Image, queue_size=10)
    pcl_pub = rospy.Publisher('kitti_point_cloud', PointCloud2, queue_size=10)
    ego_pub = rospy.Publisher('kitti_ego_car', MarkerArray, queue_size=10)
    imu_pub = rospy.Publisher('kitti_imu', Imu, queue_size=10)
    gps_pub = rospy.Publisher('kitti_gps', NavSatFix, queue_size=10)
    box3d_pub = rospy.Publisher('kitti_3d', MarkerArray, queue_size=10)

    bridge = CvBridge()

    df_tracking = read_labelfile('/home/d501//kitti_data/2011_09_26/data_tracking_label_2/training/label_02/0000.txt')
    calib = Calibration("/home/d501//kitti_data/2011_09_26/2011_09_26_drive_0005_sync/2011_09_26", from_video=True)

    rate = rospy.Rate(10)
    while not rospy.is_shutdown():
        df_tracking_frame = df_tracking[df_tracking.frame==frame]

        boxes_2d = np.array(df_tracking[df_tracking.frame==frame][['bbox_left','bbox_top','bbox_right','bbox_bottom']])
        types = np.array(df_tracking[df_tracking.frame==frame]['type'])
        boxes_3d = np.array(df_tracking_frame[['dimensions_height','dimensions_width','dimensions_length','location_x','location_y','location_z','rotation_y']])

        corners_3d_velos = []
        for box_3d in boxes_3d:
            corners_3d_cam2 = compute_3d_box_cam2(*box_3d)
            corners_3d_velo = calib.project_rect_to_velo(corners_3d_cam2.T)
            corners_3d_velos += [corners_3d_velo]

        image = read_camera(os.path.join(DATA_PATH, 'image_02/data/%010d.png'%frame))
        point_cloud = read_point_cloud(os.path.join(DATA_PATH, 'velodyne_points/data/%010d.bin'%frame))
        imu_data = read_imu(os.path.join(DATA_PATH, 'oxts/data/%010d.txt'%frame))

        publish_camera(cam_pub, bridge, image, boxes_2d, types)
        publish_point_cloud(pcl_pub, point_cloud)
        publish_ego_car(ego_pub)
        publish_imu(imu_pub, imu_data)
        publish_gps(gps_pub, imu_data)
        publish_3dbox(box3d_pub, corners_3d_velos, types)

        rospy.loginfo("published")
        rate.sleep()
        frame += 1
        frame %= 154

运行截图：

相关推荐
西猫雷婶23 分钟前
python学opencv|读取图像（二十一）使用cv2.circle()绘制圆形进阶
开发语言·python·opencv
老刘莱国瑞1 小时前
STM32 与 AS608 指纹模块的调试与应用
python·物联网·阿里云
一只敲代码的猪2 小时前
Llama 3 模型系列解析（一）
大数据·python·llama
Hello_WOAIAI2 小时前
批量将 Word 文件转换为 HTML：Python 实现指南
python·html·word
winfredzhang2 小时前
使用Python开发PPT图片提取与九宫格合并工具
python·powerpoint·提取·九宫格·照片
矩阵推荐官hy147623 小时前
短视频矩阵系统种类繁多，应该如何对比选择？
人工智能·python·矩阵·流量运营
测试19983 小时前
外包干了2年，技术退步明显....
自动化测试·软件测试·python·功能测试·测试工具·面试·职场和发展
码银3 小时前
【python】银行客户流失预测预处理部分，独热编码·标签编码·数据离散化处理·数据筛选·数据分割
开发语言·python
小木_.3 小时前
【python 逆向分析某有道翻译】分析有道翻译公开的密文内容，webpack类型，全程扣代码，最后实现接口调用翻译，仅供学习参考
javascript·python·学习·webpack·分享·逆向分析
R-sz3 小时前
14: curl#6 - “Could not resolve host: mirrorlist.centos.org； 未知的错误“
linux·python·centos
热门推荐
01【一文读懂】NTN（非地面网络）技术介绍02优化手机性能，解决卡顿问题：关闭这3个微信开关，释放内存空间03玄机平台应急响应—webshell查杀04Cube-Studio：开源大模型全链路一站式中台05基于大语言模型（LLM）的合成数据生成、策展和评估的综述06(欧拉)openEuler系统添加网卡文件配置流程、(欧拉)openEuler系统手动配置ipv6地址流程、(欧拉)openEuler系统网络管理说明0719个Web前端交互式3D JavaScript框架和库08HCIA-datacom数通题库和录播视频资料09Docker 夺命连环 15 问10Ubuntu 20.04使用Livox mid 360 测试 FAST_LIO