Semantic Classification and Navigation with Teach-and-Repeat Inspection
This repository contains semantic map algorithm for underwater vtr inspection
this script works in conjunction with yolov5ros package provided at https://github.com/mats-robotics/yolov5_ros.git and ORB_SLAM3 : https://github.com/UZ-SLAMLab/ORB_SLAM3.git During teach phase -replace groundtruth robot pose to ORB_SLAM estimated pose for objectdetector and Semanticmapmaker. -launch ORB_SLAM3 -launch yolov5 detection -launch map creator
During repeat phase -rename object_map.csv as M1.csv -rename projection.csv as C1.csv -launch ORB_SLAM3 -launch yolov5 detection -run relocatization.py
UUV_sim package used for underwater simulation check https://field-robotics-lab.github.io/dave.doc/contents/installation/Installation/ made changes to Rexrove camera sensor plugins to include depth stereo camera. -replace camera_snippet.xacro
Gazebo testing World Additional models used in the gazebo world can be downloaded from : https://drive.google.com/drive/folders/1AlWjQZcZUk3bDQ0q0uT47cm55On0YdI8?usp=share_link