Improve poincloud resolution for Euclidean cluster in long range

[badai-nguyen]

Checklist

• I've read the contribution guidelines.
• I've searched other issues and no duplicate issues were found.
• I've agreed with the maintainers that I can plan this task.

Description

Currently, the detection range is extended until 150m in front by this PR.
However, in camera_lidar_fusion mode, Euclidean cluster shows weak result of clustering from middle and long range due to of low pointcloud density or no pointcloud remained after voxel_grid and outlier_filter with the same size and threshold.

Purpose

I suggest that without downsampleing in middle and long range after ground_segmentation for Euclidean cluster input could improve the clustering result as well as camera_lidar_fusion result.

Possible approaches

1. Use some changeable size and threshold for voxel_grid and outlier_filter. It looks quite hard since pcl library modification level is needed.
2. Separate the short-range region where downsampling is applied and keep middle,long-range out of downsampling. This is more feasible IMO.
   Current pipeline around Euclidean cluster
   

Suggestion:

Definition of done

Sufficient pointclouds of compact car are remained at long-range for Euclidean_cluster.

[miursh]

closed by #2749