This file contains instructions of using Semantic DSP map with only static semantic mapping. Since our map is an ego-centric local map, a simple global mapping node is also provided. Tests were conducted with KITTI 360 dataset.
-
Download the code and build
mkdir -p semantic_map_ws/src cd semantic_map_ws/src git clone [email protected]:g-ch/mask_kpts_msgs.git git clone -b external_tracker --recursive [email protected]:g-ch/Semantic_DSP_Map.git catkin build -
Download global mapping node
cd semantic_map_ws/src [email protected]:g-ch/dsp_global_mapping.git catkin build
In include/settings/setting.h, set #define SETTING 0 and change the camera intrinsics and image size if you are using a dataset other than KITTI 360.
#if SETTING == 0 ///< KITTI_360
...
// Kitti-360 settings
constexpr float g_camera_fx_set = 552.554261;
constexpr float g_camera_fy_set = 552.554261;
constexpr float g_camera_cx_set = 682.049453;
constexpr float g_camera_cy_set = 238.769549;
constexpr int g_image_width_set = 1408;
constexpr int g_image_height_set = 376;
After the above parameters are set, run catkin build.
Information about other optional parameters can be found in Parameter Table.
- Run mapping
roslaunch semantic_dsp_map kitti360.launch
- Run global mapping. If you want to run global mapping node, please first clone and build dsp_global_map and check its readme.
roslaunch dsp_global_mapping global_mapping.launch out_file:=path_to_save_ply_global_map write_color:=1
Then feed data to topics
mask_kpts_msgs::MaskGroup /mask_group_super_glued
sensor_msgs::Image /camera_depth_image
geometry_msgs::PoseStamped /camera_pose
where in /mask_group_super_glued topic, only one mask with label "static" is required. Pixel value (one channel) in the mask stands for label id. This mask can be acquired by tools like MMSeg. The feature points are not required.