Any method that I can visualize the ORB SLAM results?
ysyyork opened this issue · 3 comments
ysyyork commented
This would be very helpful and more straightforward than pure CSV and other text file formats.
AlphaGotReal commented
You can probably visualise by passing in enable_gui=true in the ORB_SLAM3 package
bool enable_gui = false;
app.add_flag("-g,--enable_gui", enable_gui);
.
.
.
.
cv::Ptr<cv::aruco::Dictionary> aruco_dict = cv::aruco::getPredefinedDictionary(aruco_dict_id);
ORB_SLAM3::System SLAM(
vocabulary, setting,
ORB_SLAM3::System::IMU_MONOCULAR,
enable_gui, load_map, save_map, // make enable gui ---> true
aruco_dict, init_tag_id, init_tag_size
);
passing that here(03_batch_slam.py) as --enable-gui true should launch the built-in gui tool by ORB SLAM3
cmd = [
'docker',
'run',
'--rm', # delete after finish
'--volume', str(video_dir) + ':' + '/data',
'--volume', str(map_mount_source.parent) + ':' + str(map_mount_target.parent),
docker_image,
'/ORB_SLAM3/Examples/Monocular-Inertial/gopro_slam',
'--vocabulary', '/ORB_SLAM3/Vocabulary/ORBvoc.txt',
'--setting', '/ORB_SLAM3/Examples/Monocular-Inertial/gopro10_maxlens_fisheye_setting_v1_720.yaml',
'--input_video', str(video_path),
'--input_imu_json', str(json_path),
'--output_trajectory_csv', str(csv_path),
'--load_map', str(map_mount_target),
'--mask_img', str(mask_path),
'--max_lost_frames', str(max_lost_frames),
'--enable-gui', 'true' # or make it a user argument in some config file or command itself
]
but before that you also need to enable gui in docker by running the xhost +
or xhost +docker
command in the particular terminal window