Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Delete several packages/implementations in perception directory #4165

Open
xmfcx opened this issue Feb 9, 2024 Discussed in #4136 · 2 comments
Open

Delete several packages/implementations in perception directory #4165

xmfcx opened this issue Feb 9, 2024 Discussed in #4136 · 2 comments
Assignees
Labels
component:perception Advanced sensor data processing and environment understanding.

Comments

@xmfcx
Copy link
Contributor

xmfcx commented Feb 9, 2024

Discussed in https://github.com/orgs/autowarefoundation/discussions/4136

Originally posted by kminoda January 31, 2024

Introduction

I would like to delete several packages/implementations in perception, listed below:

  • heatmap_visualizer
  • lidar_centerpoint/single_inference
  • traffic_light_ssd_fine_detector
  • tensorrt_yolo
  • front_vehicle_velocity_estimator

Justification

heatmap_visualizer

https://github.com/autowarefoundation/autoware.universe/tree/main/perception/heatmap_visualizer

This is a script which is not intended to run on runtime. I have confirmed with the author of this script and agreed that we can delete this package.

lidar_centerpoint/single_inference

https://github.com/autowarefoundation/autoware.universe/blob/main/perception/lidar_centerpoint/launch/single_inference_lidar_centerpoint.launch.xml

Single inference script for CenterPoint has been added 2 years ago starting from this issue: https://github.com/orgs/autowarefoundation/discussions/2755. However, the script is not intended to run on runtime but rather at a debug time, which makes it difficult for us to apply CI on this implementation. If there is no one currently using this script, we would like to delete this launch file and some related files as well.

traffic_light_ssd_fine_detector

https://github.com/autowarefoundation/autoware.universe/tree/main/perception/traffic_light_ssd_fine_detector

This package is primarily implemented for traffic light detection using SSD. However, with the recent development, current Autoware uses traffic_light_fine_detector as a default option, outperforming SSD detector in most cases.
Please let us know if you are currently using this package. Also, let us know if it is difficult for you to make a transition to traffic_light_fine_detector.

tensorrt_yolo

https://github.com/autowarefoundation/autoware.universe/tree/main/perception/tensorrt_yolo

This package performs 2D detection on camera images to achieve camera-LiDAR fusion. This is based on YOLO, but recently TIER IV has developed YOLOX version (tensorrt_yolox) which outperforms this package in most of the cases.

front_vehicle_velocity_estimator

https://github.com/autowarefoundation/autoware.universe/tree/main/perception/front_vehicle_velocity_estimator

This package is currently not used in default Autoware implementation. I confirmed with the author of this script and agreed that we can delete this package.

Process of deleting these files

We will have a survey period for 2 weeks from today until 16th of February. Please comment if you have any oppositions or opinions on either of the listed files.

After this period, we are going to delete all the files listed above as long as there is no disagreements or concerns raised in this Discussion.

Related links

@xmfcx
Copy link
Contributor Author

xmfcx commented Feb 21, 2024

@kminoda If there are any other related PRs, could you link them to this issue?

If all are completed, I would like to close this issue.

It's better to link issues rather than discussions in PRs.

Because when the issues are mentioned in PRs, they are back-linked automatically under the issue itself.

@kminoda
Copy link
Contributor

kminoda commented Feb 22, 2024

It's almost done, the only ToDs left is the lidar_centerpoint/single_inference. Will do that when I have time (or also appreciated if anyone could do that instead 👍 )

Other PRs are all properly linked. Thanks! @xmfcx

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
component:perception Advanced sensor data processing and environment understanding.
Projects
None yet
Development

No branches or pull requests

2 participants