This repository contains implementation of the visualisation tools mentioned in the scope of Issue #35 of the BenchBot Software Stack.
Tools allow to visualise scene semantic map and overlay two semantic maps (e.g. ground-truth vs predicted) as illustrated in the pictures below:
Ground-truth semantic map | Ground-truth semantic map vs semantic map built using VoteNet |
---|---|
![]() |
![]() |
Repository structure:
ssu_map3d_visualization/
- data/
| - miniroom_1.json # ground-truth objects bounding boxes
| - my_semantic_slam_0.json # objects bounding boxes pridicted by VoteNet
- img/ # example semantic map visualisations
| - miniroom1_ground_truth.png
| - miniroom1_ground_truth_vs_predicted.png
- vis_tools/ # python package with visualisation tools
| - __init__.py
| - tools.py
- gt_map_vis.py # example how to visualise one semantic map
- gt_vs_pr_map_vis.py # example how to overlay two semantic maps