using the KITTI dataset are not as good as those in the paper.

[lijing137]

Hi，Thank you very much for making this excellent work available. However, I encountered some issues while using it. The results obtained using the KITTI dataset are not as good as those presented in the paper. I hope you can provide some guidance.

Result(former is mine, latter is paper)

platform
ubuntu20.04;Ros1

dataset
i use the kitti2bag tool generate a *.bag file (kitti_2011_09_26_drive_0039_synced.bag)
https://www.cvlibs.net/datasets/kitti/raw_data.php?type=residential

parameter
parameter use the default /config/KITTI.yaml

voxel_size: 0.05
sdf_trunc: 0.1
space_carving: False

fill_holes: True
min_weight: 3.0

apply_pose: True
preprocess: True
min_range: 1.0
max_range: 75.0
pcl_topic: "/kitti/velo/pointcloud"

use_tf_transforms: true
parent_frame: "world"
child_frame: "base_link"
timestamp_tolerance_ns: 10000000

I hope you can provide some guidance. Thank you very much！