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

feat: add point filters designed based on overlapping LiDAR regions #401

Draft
wants to merge 1 commit into
base: tier4/universe
Choose a base branch
from

Conversation

mojomex
Copy link
Contributor

@mojomex mojomex commented Mar 6, 2025

This PR introduces point filters designed with a novel method:

  1. projecting each LiDAR into each other's FoV
  2. visualizing for each LiDAR over azimuth-elevation
    a. the effective angular resolution
    b. the number of overlapping LiDARs in every region
  3. using that information to downsample more in regions of high overlap
  4. downsample additionally in azimuth-elevation regions that have a given maximum distance (e.g. always hit the ground a few meters away)

The first results can be seen here:

point_filters_full point_filters_one_third `overlap_based
image image image
500k points 220k points 150k points

These filters are not yet final and are more of a proof of concept.
With the same methodology and proper testing for low obstacles and far-away cars, production-ready filters can be designed with relative ease.

See TIER IV INTERNAL for details on the design process.

Use this tool to reproduce this methodology.

See [TIER IV INTERNAL](https://tier4.atlassian.net/wiki/spaces/CT/pages/3569713951/S+P+nebula+downsample#Overlap-Based-Settings) for details on the design process.

Use [this tool](https://github.com/tier4/project_pointclouds) to reproduce this methodology.

Signed-off-by: Max SCHMELLER <max.schmeller@tier4.jp>
@mojomex mojomex self-assigned this Mar 6, 2025
@YoshiRi YoshiRi requested a review from kotaro-hihara March 10, 2025 08:05
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant