fix(euclidean_cluster): fix euclidean cluster params

[beginningfan]

Description

Fix bug 2 about max_cluster_size_ related in here
The bug about min_cluster_size is a mistake and it should not to be changed

Tests performed

use a large max_cluster_size_ will not take too much CPU and memory

Effects on system behavior

Not applicable.

Interface changes

Pre-review checklist for the PR author

The PR author must check the checkboxes below when creating the PR.

• I've confirmed the contribution guidelines.
• The PR follows the pull request guidelines.

In-review checklist for the PR reviewers

The PR reviewers must check the checkboxes below before approval.

• The PR follows the pull request guidelines.

Post-review checklist for the PR author

The PR author must check the checkboxes below before merging.

• There are no open discussions or they are tracked via tickets.

After all checkboxes are checked, anyone who has write access can merge the PR.

[yukkysaito]

@badai-nguyen Can I ask you to review this?

[badai-nguyen]

@beginningfan Thank you for your PR and raising issue.

• Related to min_cluster_size I commented here Bugs about euclidean_cluster #7587 (comment)
• Related to max_cluster_size, since the voxel_grid_filter was applied to downsized input pointcloud before applying euclidean_cluster filter so that the actual output cluster size after reversing the voxel_grid_filter could be bigger than cluster.indices.size(). It means that applying your proposal might over_filtering out some clusters which has actual cluster_size < max_cluster_size_. That is also the reason why min_cluster_size for pcl_euclidean_cluster_filter is set as 1 and rechecked at later stage.

Could you check the issue that you raised again, and let me known if you have further comment.

[beginningfan]

@badai-nguyen thanks for your reply.
Related to min_cluster_size, I misunderstood this, so I did not modify min_cluster_size in this PR.
Related to max_cluster_size, I understand your approach. I first allocate space according to the downsized cluster size, and then resize it if there is insufficient space when push_back. Because mostly we cannot determine the maximum value of clusters, so we set a larger number to ensure that clusters are not lost. However, using the maximum value to allocate space will cause excessive CPU and memory usage.This method will not cause problems in previous versions.

[badai-nguyen]

using the maximum value to allocate space will cause excessive CPU and memory usage.

@beginningfan I agree with you.
However, I think decide that the max size of output cluster (cluster_data_size) to be doubled of temporary_cluster is not reasonable. (runtime_error might be occurred here for some high dense pointcloud)

Suggested change

	.data.resize(temporary_clusters.at(map[index]).data.size() * 2);
	.data.resize(max_cluster_size_ * point_step);

Or we can increment cluster_data_size and check it repeatedly.

    if (cluster_data_size >= temporary_clusters.at(map[index]).data.size()) {
        temporary_clusters.at(map[index]).data.resize(cluster_data_size + point_step * 10);
      }
      std::memcpy(
        &temporary_clusters.at(map[index]).data[cluster_data_size],
        &pointcloud_msg->data[i * point_step], point_step);
      cluster_data_size += point_step;
    }

[beginningfan]

Here is a for loop that reallocates twice the space every time the space is full. When the space is full, it will twice the space again.I have tested using a ouster64, and the dense pointcloud is sometimes 50 times larger than the downsampled one. Even so, enough space can be allocated quickly.
If you use max_cluster_size_, it might as well just use the original code. If we add 10 points each time, this will result in multiple copies of the vector data.

[badai-nguyen]

@beginningfan I understood. I have checked and it worked well. Thank you for your improvement.

[codecov]

Codecov Report

Attention: Patch coverage is 0% with 4 lines in your changes missing coverage. Please review.

Project coverage is 2.56%. Comparing base (3d849e9) to head (69483aa).
Report is 548 commits behind head on main.

Files	Patch %	Lines
...cluster/lib/voxel_grid_based_euclidean_cluster.cpp	0.00%	4 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #7662       +/-   ##
==========================================
- Coverage   15.09%   2.56%   -12.54%     
==========================================
  Files        1967     201     -1766     
  Lines      135941   10420   -125521     
  Branches    42122    1744    -40378     
==========================================
- Hits        20520     267    -20253     
+ Misses      92700   10097    -82603     
+ Partials    22721      56    -22665     

Flag	Coverage Δ
differential	2.56% <0.00%> (?)
total	?

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[badai-nguyen]

LGTM

[github-actions]

Thank you for contributing to the Autoware project!

🚧 If your pull request is in progress, switch it to draft mode.

Please ensure:

• You've checked our contribution guidelines.
• Your PR follows our pull request guidelines.
• All required CI checks pass before marking the PR ready for review.

[beginningfan]

@mitsudome-r The reviewer of this PR has approved it. Can it be merged now?