不过DS-SLAM通过查看在语义分割图上该像素的灰度值来判断该像素点是否有语义信息,如果有语义信息就将语义分割图片的颜色赋给点云,颜色信息从2D映射到3D;否则点云的信息就是正常RGBD的颜色,代码如下: pcl::PointCloud< PointCloudMapping::PointT >::Ptr PointCloudMapping::generatePoint
DS-SLAM could also provide semantic presentation of the octo-tree map.DS-SLAM is a optimized SLAM system based on the famous ORB-SLAM2 (fromhttps://github.com/raulmur/ORB_SLAM2andhttps://github.com/gaoxiang12/ORBSLAM2_with_pointcloud_map, thanks for Raul's, Gao Xiang's and SegNet's...
roslaunch rds_slam tum_segnet_walk_static.launch roslaunch rds_slam tum_segnet_walk_xyz.launch Dataset Let docker access your dataset if you use docker vim docker/.env DATASET_DIR=/.../data/Dataset TUM dataset RDS-SLAM evaluated using the TUM dataset Please download the TUM dataset sequence...