AutoLidarPerception / kitti_ros

A ROS-based player to replay KiTTI dataset. http://www.cvlibs.net/datasets/kitti/
http://www.cvlibs.net/datasets/kitti/
30 stars 7 forks source link

[keyboard_file] Three is no image in the rviz. #26

Closed yileichen123 closed 3 years ago

yileichen123 commented 5 years ago

I had download the kitti data and I had all the requirements.But there is no images and points on the screen. And there is no error.It just show and pause as the following: [ 2011-09-26 13:04:32.335337 ] # of Point Clouds: 493588 ########### [INFO] Show image: 0000000000.png

And the url: ROS_MASTER_URI=http://localhost:11311, I can't cotact on the website

maxwellfrom commented 5 years ago

I met the same problem, help! @Durant35 Thanks!

maxwellfrom commented 5 years ago

Have you solved the problem? @yileichen123

Durant35 commented 5 years ago

Have you press the keyboard to control the data replay?

maxwellfrom commented 5 years ago

Yes, I have tried to use keyboard, but it didn't work. @Durant35

maxwellfrom commented 5 years ago

just stoped at [ 2011-09-26 13:04:32.335337 ] # of Point Clouds: 493588 ########### [INFO] Show image: 0000000000.png

Durant35 commented 5 years ago

Could you please upload a screenshot?

maxwellfrom commented 5 years ago

screenshot from 2019-01-11 20-43-11

maxwellfrom commented 5 years ago

The same situation appeared when I used 'quickstart '

maxwellfrom commented 5 years ago

This is the rviz screenshot : screenshot from 2019-01-11 20-54-08

Durant35 commented 5 years ago

Have you try?

image

maxwellfrom commented 5 years ago

yes, I have tried that.

Durant35 commented 5 years ago

I have just followed the README, and everything goes well. image

Durant35 commented 5 years ago

@maxwellfrom Maybe you can try the latest version and find is this issue still persistent?

maxwellfrom commented 5 years ago

I tried your latest version, and it was same as before.

maxwellfrom commented 5 years ago

When I use the keyboard to control, is there anything needs to be noticed?

maxwellfrom commented 5 years ago

There is a warn is rviz, I wonder if it is because of that warn. screenshot from 2019-01-13 20-31-11

Durant35 commented 5 years ago

When you click SPACE,data replay,and you will see data info frame by frame.

maxwellfrom commented 5 years ago

I click SPACE there is no response, What is your terminal output when you run ./quickstart.sh, is that like this: screenshot from 2019-01-13 20-42-34

Durant35 commented 5 years ago

what error?roslaunch will exit,means something wrong!

maxwellfrom commented 5 years ago

I run the roscore firstly, that problem disappeared : screenshot from 2019-01-13 21-02-14

so , it may be not because of that.

maxwellfrom commented 5 years ago

Could you please show me your whole rviz screenshot? Thanks!

Durant35 commented 5 years ago
maxwellfrom commented 5 years ago

There is difference between your rviz and mine: screenshot from 2019-01-13 20-31-11 , I just followed your README step by step , I really don't why .

Durant35 commented 5 years ago

I think that's not the problem... Actually it's due to that you never successfully launch the kitti_player or never replay any data using that kitti_player.

Latest version is just for looking beautiful, as following: image

Durant35 commented 5 years ago

@maxwellfrom @yileichen123 I think the only problem is the keyboard callback function couldn't be triggered, and the error information will be shown on your terminal, supported by ROS's roslaunch.

If you use any IDE, like VS code above, I'm afraid that I can't help more, sorry.

Durant35 commented 5 years ago

image

image

maxwellfrom commented 5 years ago

What is your Ubuntu version, my version is 16.04. I wonder if it is caused by the difference between Ubuntu versions.

Durant35 commented 5 years ago

Ubuntu 16.04

ROS Kinetic

maxwellfrom commented 5 years ago

The problem still exists, but you have done a good job. Do you have any plan to join a self-driving company? I am a bachelor student of USTC and I'm now an intern at a self-driving company, I can recommend you to my boss.

Durant35 commented 5 years ago

@maxwellfrom Never mind to contact me, chenshj35@mail2.sysu.edu.cn, hope to solve your problem...

Durant35 commented 5 years ago

@maxwellfrom I met the same problem on my PC, with different input devices. Then I try the following command and find that the Keyboard is not the same as default /dev/input/event3 on my laptop.

$ cat /var/log/Xorg.0.log | grep Keyboard
[     4.555] (II) config/udev: Adding input device Logitech USB Keyboard (/dev/input/event4)
[     4.555] (**) Logitech USB Keyboard: Applying InputClass "evdev keyboard catchall"
[     4.555] (II) Using input driver 'evdev' for 'Logitech USB Keyboard'
[     4.555] (**) Logitech USB Keyboard: always reports core events
[     4.555] (**) evdev: Logitech USB Keyboard: Device: "/dev/input/event4"
[     4.555] (--) evdev: Logitech USB Keyboard: Vendor 0x46d Product 0xc31c
[     4.555] (--) evdev: Logitech USB Keyboard: Found keys
[     4.555] (II) evdev: Logitech USB Keyboard: Configuring as keyboard
[     4.555] (II) XINPUT: Adding extended input device "Logitech USB Keyboard" (type: KEYBOARD, id 10)
[     4.555] (II) config/udev: Adding input device Logitech USB Keyboard (/dev/input/event5)
[     4.555] (**) Logitech USB Keyboard: Applying InputClass "evdev keyboard catchall"
[     4.555] (II) Using input driver 'evdev' for 'Logitech USB Keyboard'
[     4.555] (**) Logitech USB Keyboard: always reports core events
[     4.555] (**) evdev: Logitech USB Keyboard: Device: "/dev/input/event5"
[     4.555] (--) evdev: Logitech USB Keyboard: Vendor 0x46d Product 0xc31c
[     4.555] (--) evdev: Logitech USB Keyboard: Found absolute axes
[     4.555] (II) evdev: Logitech USB Keyboard: Forcing absolute x/y axes to exist.
[     4.555] (--) evdev: Logitech USB Keyboard: Found keys
[     4.555] (II) evdev: Logitech USB Keyboard: Forcing relative x/y axes to exist.
[     4.555] (II) evdev: Logitech USB Keyboard: Configuring as mouse
[     4.555] (II) evdev: Logitech USB Keyboard: Configuring as keyboard
[     4.555] (II) XINPUT: Adding extended input device "Logitech USB Keyboard" (type: KEYBOARD, id 11)
[     4.555] (II) evdev: Logitech USB Keyboard: initialized for absolute axes.
[     4.555] (**) Logitech USB Keyboard: (accel) keeping acceleration scheme 1
[     4.555] (**) Logitech USB Keyboard: (accel) acceleration profile 0
[     4.555] (**) Logitech USB Keyboard: (accel) acceleration factor: 2.000
[     4.555] (**) Logitech USB Keyboard: (accel) acceleration threshold: 4

And finally it works after trying both input devices listed above...

$ roslaunch kitti_ros kitti_player.launch kitti_data_path:=/home/gary/Datasets/KiTTI/2011_09_26/2011_09_26_drive_0005_sync keyboard_file:=/dev/input/event4
... logging to /home/gary/.ros/log/a530e1d2-18bd-11e9-b531-309c23a01982/roslaunch-gary-MS-7A71-12587.log
Checking log directory for disk usage. This may take awhile.
Press Ctrl-C to interrupt
Done checking log file disk usage. Usage is <1GB.

WARNING: Package name "rosAD" does not follow the naming conventions. It should start with a lower case letter and only contain lower case letters, digits, underscores, and dashes.
started roslaunch server http://gary-MS-7A71:42192/

SUMMARY
========

PARAMETERS
 * /kitti_player/care_objects: ['Car', 'Van', 'T...
 * /kitti_player/filter_by_camera_angle: True
 * /kitti_player/fps: 10
 * /kitti_player/imu_frame: imu
 * /kitti_player/keyboard_file: /dev/input/event4
 * /kitti_player/kitti_data_path: /home/gary/Datase...
 * /kitti_player/mode: play
 * /kitti_player/vel_frame: velodyne
 * /kitti_player/world_frame: world
 * /rosdistro: kinetic
 * /rosversion: 1.12.13

NODES
  /
    kitti_player (kitti_ros/kitti_player.py)

ROS_MASTER_URI=http://localhost:11311

process[kitti_player-1]: started with pid [12604]
parsing tracklet file /home/gary/Datasets/KiTTI/2011_09_26/2011_09_26_drive_0005_sync/tracklet_labels.xml
file contains 15 tracklets
loaded 15 tracklets

[ 2011-09-26 13:04:32.335337 ] # of Point Clouds: 493588
###########
[INFO] Show image:  0000000000.png

[ 2011-09-26 13:04:32.438625 ] # of Point Clouds: 494732
 ###########
[INFO] Show image:  0000000001.png

[ 2011-09-26 13:04:32.541951 ] # of Point Clouds: 495132
###########
[INFO] Show image:  0000000002.png
...
Durant35 commented 5 years ago

Seems that Keyboard is not allowed during teamviewer with @maxwellfrom, only works by host control.

Durant35 commented 5 years ago

Move to LidarPerception/kitti_ros_tmp#4

wdxpython commented 5 years ago

Seems that Keyboard is not allowed during teamviewer with @maxwellfrom, only works by host control.

大神,以你上面的图为例,cat /var/log/Xorg.0.log | grep Keyboard之后显示有"/dev/input/event4"和"/dev/input/event5" 两个 sudo chmod 777 /dev/input/event4 event5 ?这样?

Durant35 commented 5 years ago

@wdxpython Please pay attention to the keyboard_file parameter as following (also mentioned above), details refer to https://github.com/LidarPerception/kitti_ros#parameters

$ roslaunch kitti_ros kitti_player.launch kitti_data_path:=/home/gary/Datasets/KiTTI/2011_09_26/2011_09_26_drive_0005_sync keyboard_file:=/dev/input/event4
Durant35 commented 5 years ago

or you can directly change the default value for keyboard_file in related launch file.

wdxpython commented 5 years ago

or you can directly change the default value for keyboard_file in related launch file.

Thank you very much for your patience. I have solved this problem and I will continue to follow your code. I saw that you used the variable-threshold European segmentation method in region_euclidean_segmenter.cpp. I have also experimented with it myself, but the real-time performance is not very good. In your reference, I saw the paper on the Fast segmentation of 3d point clouds: A paradigm on lidar data for autonomous vehicle applications. The ground segmentation part of this paper, I can understand, segment the obstacles of the road surface. In theory, I also know how to do it, but there is a problem in code implementation. I saw someone reproduce the code https://github.com/VincentCheungM/Run_based_segmentation, which has some problems in the boundingbox part. I modified it and found that the result was too bad. Because of my poor ability to analyze the code, I analyzed the problem and found no problem. . I hope that in the future, the part that divides the obstacles on the road can also be integrated into your library.