Multiclass Semantic Segmentation

[VAM7686]

Description

This PR does the following:

• Implements the Multiclass Segmentation model

Fixes #747

Testing steps

1. Make sure to have the model downloaded and placed in the igvc_sandbox/models folder
   • (You can either train your own model or download the one located in Shared Drives > RoboNav > Software > model_training > full_model4_resized > checkpoints > best.pth)
2. Run roscore
3. Run the rosbag file with something like: rosbag play ~/path/to/file/bag_file.bag -l
   • Try using bags from the 2021-2022 dataset such as the swervi_parking_lot_2022-05-27-19-46-11.bag
4. Run the multiclass with: roslaunch igvc_perception multiclass_segmentation.launch
5. Launch RVIZ and add the /cam/(center/right/left)/segmented/image... topic
6. Make sure that semantic segmentation output is visualized on the aforementioned topics
   • /cam/center/segmented/image, /cam/right/segmented/image, /cam/left/segmented/image, etc

Expected Output (old version on grass course)

• https://www.dropbox.com/s/yuv4jou4ch4uqla/multiclassSegmentation.webm?dl=0

Self Checklist

• [x] I have formatted my code using make format
• [x] I have tested that the new behaviour works

[pep8speaks]

Hello @VAM7686! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found:

There are currently no PEP 8 issues detected in this Pull Request. Cheers! :beers:

Comment last updated at 2022-06-03 21:17:09 UTC