Following the instructions and using the docker image included in this repo, you should be able to run the tracking and mapping algorithm of OpenVSLAM on a ros-noetic-base system.
You might need to install docker on your EC2 machine.
The docker images that we are going to use are available here.
sudo docker pull debbieliang/ros_open:open_ros_log
sudo docker pull debbieliang/ros_open:open_ros_server
sudo docker run --network host --rm -it debbieliang/ros_open:open_ros_log
cd ../ros
source /openvslam/ros/devel/setup.bash
sudo docker run --network host --rm -it debbieliang/ros_open:open_ros_log
cd ../ros
source /openvslam/ros/devel/setup.bash
sudo docker run --network host --rm -it debbieliang/ros_open:open_ros_log
cd ../..
roscore
sudo docker run --network host --rm -it debbieliang/ros_open:open_ros_server
You should see
WebSocket: listening on *:3000
HTTP server: listening on *:3001
Now we've finished the setup.
As a subscriber, run rosrun openvslam run_slam -v /openvslam/build/orb_vocab/orb_vocab.dbow2 -c /openvslam/build/aist_living_lab_1/config.yaml --map-db map.msg
As a publisher, run rosrun publisher video -m /openvslam/build/aist_living_lab_1/video.mp4
You should be able to see
[2021-02-28 18:49:15.924] [I] config file loaded: /openvslam/build/aist_living_lab_1/config.yaml
___ __ _____ _ _ __ __
/ _ \ _ __ ___ _ _\ \ / / __| | /_\ | \/ |
| (_) | '_ \/ -_) ' \\ V /\__ \ |__ / _ \| |\/| |
\___/| .__/\___|_||_|\_/ |___/____/_/ \_\_| |_|
|_|
Copyright (C) 2019,
National Institute of Advanced Industrial Science and Technology (AIST)
All rights reserved.
This is free software,
and you are welcome to redistribute it under certain conditions.
See the LICENSE file.
Camera Configuration:
- name: RICOH THETA S 960
- setup: Monocular
- fps: 30
- cols: 1920
- rows: 960
- color: RGB
- model: Equirectangular
ORB Configuration:
- number of keypoints: 2000
- scale factor: 1.2
- number of levels: 8
- initial fast threshold: 20
- minimum fast threshold: 7
- mask rectangles:
- [0, 1, 0, 0.1]
- [0, 1, 0.84, 1]
- [0, 0.2, 0.7, 1]
- [0.8, 1, 0.7, 1]
[2021-02-28 18:49:15.924] [I] loading ORB vocabulary: /openvslam/build/orb_vocab/orb_vocab.dbow2
[2021-02-28 18:49:16.396] [I] startup SLAM system
[2021-02-28 18:49:16.396] [I] start mapping module
[2021-02-28 18:49:16.396] [I] start global optimization module
[2021-02-28 18:49:16] [connect] Successful connection
[2021-02-28 18:49:16] [connect] WebSocket Connection 127.0.0.1:3000 v-2 "WebSocket++/0.8.1" /socket.io/?EIO=4&transport=websocket&t=1614538156 101
[2021-02-28 18:49:16.399] [I] connected to server
[2021-02-28 18:49:16.782] [I] initialization succeeded with E
[2021-02-28 18:49:16.815] [I] new map created with 113 points: frame 0 - frame 1
[2021-02-28 18:49:23.966] [I] tracking lost within 5 sec after initialization
[2021-02-28 18:49:23.969] [I] resetting system
[2021-02-28 18:49:23.969] [I] reset mapping module
[2021-02-28 18:49:23.975] [I] reset global optimization module
[2021-02-28 18:49:23.975] [I] clear BoW database
[2021-02-28 18:49:23.988] [I] clear map database
[2021-02-28 18:49:24.102] [I] initialization succeeded with E
[2021-02-28 18:49:24.138] [I] new map created with 187 points: frame 0 - frame 1
[2021-02-28 18:50:09.496] [I] detect loop: keyframe 28 - keyframe 131
[2021-02-28 18:50:09.500] [I] pause mapping module
[2021-02-28 18:50:09.919] [I] resume mapping module
[2021-02-28 18:50:09.919] [I] start loop bundle adjustment
[2021-02-28 18:50:11.378] [I] finish loop bundle adjustment
[2021-02-28 18:50:11.378] [I] updating the map with pose propagation
[2021-02-28 18:50:11.475] [I] pause mapping module
[2021-02-28 18:50:11.507] [I] resume mapping module
[2021-02-28 18:50:11.507] [I] updated the map
On terminal A. Wait for ~3 min for the algorithm to finish, use Control + C to terminate the process. A map.msg file should be saved in the /openvslam/ros directory.
If you want to see the tracking times, add --eval-log flag when running the rosrun openvslam run_slam command in terminal A. After terminating the process, a file called track_times.txt that logs each time period of frame tracking. While the process is running, this information is also writen to a file called track_times_all.txt.
As a subscriber, run rosrun openvslam run_localization -v /openvslam/build/orb_vocab/orb_vocab.dbow2 -c /openvslam/build/aist_living_lab_1/config.yaml --map-db map.msg
As a publisher, run rosrun publisher video -m /openvslam/build/aist_living_lab_1/video.mp4
You should be able to see
[2021-02-28 19:02:15.821] [I] config file loaded: /openvslam/build/aist_living_lab_1/config.yaml
___ __ _____ _ _ __ __
/ _ \ _ __ ___ _ _\ \ / / __| | /_\ | \/ |
| (_) | '_ \/ -_) ' \\ V /\__ \ |__ / _ \| |\/| |
\___/| .__/\___|_||_|\_/ |___/____/_/ \_\_| |_|
|_|
Copyright (C) 2019,
National Institute of Advanced Industrial Science and Technology (AIST)
All rights reserved.
This is free software,
and you are welcome to redistribute it under certain conditions.
See the LICENSE file.
Camera Configuration:
- name: RICOH THETA S 960
- setup: Monocular
- fps: 30
- cols: 1920
- rows: 960
- color: RGB
- model: Equirectangular
ORB Configuration:
- number of keypoints: 2000
- scale factor: 1.2
- number of levels: 8
- initial fast threshold: 20
- minimum fast threshold: 7
- mask rectangles:
- [0, 1, 0, 0.1]
- [0, 1, 0.84, 1]
- [0, 0.2, 0.7, 1]
- [0.8, 1, 0.7, 1]
[2021-02-28 19:02:15.822] [I] loading ORB vocabulary: /openvslam/build/orb_vocab/orb_vocab.dbow2
[2021-02-28 19:02:16.307] [I] clear map database
[2021-02-28 19:02:16.308] [I] clear BoW database
[2021-02-28 19:02:16.308] [I] load the MessagePack file of database from map.msg
[2021-02-28 19:02:18.063] [I] decoding 1 camera(s) to load
[2021-02-28 19:02:18.063] [I] load the tracking camera "RICOH THETA S 960" from JSON
[2021-02-28 19:02:18.411] [I] decoding 146 keyframes to load
[2021-02-28 19:02:21.504] [I] decoding 9963 landmarks to load
[2021-02-28 19:02:21.520] [I] registering essential graph
[2021-02-28 19:02:21.987] [I] registering keyframe-landmark association
[2021-02-28 19:02:22.495] [I] updating covisibility graph
[2021-02-28 19:02:22.536] [I] updating landmark geometry
[2021-02-28 19:02:23.207] [I] startup SLAM system
[2021-02-28 19:02:23.209] [I] start mapping module
[2021-02-28 19:02:23.209] [I] start global optimization module
[2021-02-28 19:02:23.222] [I] pause mapping module
[2021-02-28 19:02:23] [connect] Successful connection
[2021-02-28 19:02:23] [connect] WebSocket Connection 127.0.0.1:3000 v-2 "WebSocket++/0.8.1" /socket.io/?EIO=4&transport=websocket&t=1614538943 101
[2021-02-28 19:02:23.268] [I] connected to server
[2021-02-28 19:02:23.804] [I] relocalization succeeded
[2021-02-28 19:02:31.424] [I] tracking lost: frame 740
[2021-02-28 19:02:31.508] [I] relocalization succeeded
On terminal A. Wait for ~3 min for the algorithm to finish, use Control + C to terminate the process.
rosrun openvslam run_slam -v /openvslam/build/orb_vocab/orb_vocab.dbow2 -c /openvslam/build/fisheye_datasets/aist_entrance_hall_1/config.yaml --map-db map.msg
rosrun publisher video -m /openvslam/build/fisheye_datasets/aist_entrance_hall_1/video.mp4
rosrun openvslam run_slam -v /openvslam/build/orb_vocab/orb_vocab.dbow2 -c /openvslam/build/fisheye_datasets/nu_eng2_corridor_1/config.yaml --map-db map.msg
rosrun publisher video -m /openvslam/build/fisheye_datasets/nu_eng2_corridor_1/video.mp4
git clone https://github.com/Debbieliang9/openvslam.git
cd ~/openvslam
docker build -t openvslam-ros .
docker run --network host --rm -it openvslam-ros
The Dockerfile used here is a combination of ros-base and openvslam-socket. Now you should be in the ~/openvslam/build file
# download an ORB vocabulary from Google Drive
FILE_ID="1wUPb328th8bUqhOk-i8xllt5mgRW4n84"
curl -sc /tmp/cookie "https://drive.google.com/uc?export=download&id=${FILE_ID}" > /dev/null
CODE="$(awk '/_warning_/ {print $NF}' /tmp/cookie)"
curl -sLb /tmp/cookie "https://drive.google.com/uc?export=download&confirm=${CODE}&id=${FILE_ID}" -o orb_vocab.zip
unzip orb_vocab.zip
# download a sample dataset from Google Drive
FILE_ID="1d8kADKWBptEqTF7jEVhKatBEdN7g0ikY"
curl -sc /tmp/cookie "https://drive.google.com/uc?export=download&id=${FILE_ID}" > /dev/null
CODE="$(awk '/_warning_/ {print $NF}' /tmp/cookie)"
curl -sLb /tmp/cookie "https://drive.google.com/uc?export=download&confirm=${CODE}&id=${FILE_ID}" -o aist_living_lab_1.zip
unzip aist_living_lab_1.zip
# download a sample dataset from Google Drive
FILE_ID="1TVf2D2QvMZPHsFoTb7HNxbXclPoFMGLX"
curl -sc /tmp/cookie "https://drive.google.com/uc?export=download&id=${FILE_ID}" > /dev/null
CODE="$(awk '/_warning_/ {print $NF}' /tmp/cookie)"
curl -sLb /tmp/cookie "https://drive.google.com/uc?export=download&confirm=${CODE}&id=${FILE_ID}" -o aist_living_lab_2.zip
unzip aist_living_lab_2.zip
(This is downloading the dataset from the OpenVSLAM tutorial)
cd ~/openvslam/viewer
docker build -t openvslam-server .
docker run --rm -it --name openvslam-server -p 3001:3001 openvslam-server
You should see
WebSocket: listening on *:3000
HTTP server: listening on *:3001
You can also run it with docker run --network host --rm -it --name openvslam-server openvslam-server and you don't have to do the work in terminal C. However, this might affect your visualization and your ability to terminate the program and save the map.
docker inspect openvslam-server | grep -m 1 \"IPAddress\" | sed 's/ //g' | sed 's/,//g'
You should see terminal C returning an IPAdress. For example, if you see
"IPAddress": "172.17.0.2"
Please go to ~/openvslam/build/aist_living_lab_1 on terminal A and append
SocketPublisher.server_uri: "http://172.17.0.2:3000"
to the config.yaml file. Note that this number can change while your container is running. If your worked connection does not work anymore, run the docker inspect command again and change the ymal file.
You may need
apt-get update
apt-get Install vim
to install vim.
visit http://localhost:3001/
In the ~/openvslam/build directory, run ./run_video_slam -v ./orb_vocab/orb_vocab.dbow2 -m ./aist_living_lab_1/video.mp4 -c ./aist_living_lab_1/config.yaml --frame-skip 3 --no-sleep --map-db map.msg.
You should be able to see the visualization on http://localhost:3001/.
Click the [Terminate] button to close the viewer. You can find map.msg in the current directory of terminal A.
(The instructions above are for MacOS, if you would like to have linux instructions, please see the openvslam Docker tutorial as your
Reference)
Reference Finish all the steps in "On your local machine without ROS" except the last one on terminal A.
Started by docker run --network host --rm -it open_ros.
cd ../..
roscore
Run
apt-get install libboost-all-dev
apt update -y
apt install ros-${ROS_DISTRO}-image-transport
cd /openvslam/ros
git clone --branch ${ROS_DISTRO} --depth 1 https://github.com/ros-perception/vision_opencv.git
cp -r vision_opencv/cv_bridge src/
rm -rf vision_opencv
catkin_make \
-DBUILD_WITH_MARCH_NATIVE=ON \
-DUSE_PANGOLIN_VIEWER=OFF \
-DUSE_SOCKET_PUBLISHER=ON \
-DUSE_STACK_TRACE_LOGGER=ON \
-DBOW_FRAMEWORK=DBoW2
source /openvslam/ros/devel/setup.bash
Repeat what terminal A just did.
For tracking and mapping, as a subscriber, run rosrun openvslam run_slam -v /openvslam/build/orb_vocab/orb_vocab.dbow2 -c /openvslam/build/aist_living_lab_1/config.yaml --map-db map.msg
As a publisher, run rosrun publisher video -m /openvslam/build/aist_living_lab_1/video.mp4
You should be able to see the visualization on http://localhost:3001/.
Click the [Terminate] button to close the viewer. You can find map.msg in the current directory of terminal A.
As a subscriber, run rosrun openvslam run_localization -v /openvslam/build/orb_vocab/orb_vocab.dbow2 -c /openvslam/build/aist_living_lab_1/config.yaml --map-db map.msg
As a publisher, run rosrun publisher video -m /openvslam/build/aist_living_lab_1/video.mp4