1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
|
# This is a config file for the `TICPSequential.cpp` Example.
# When voxel_size is -1, no downsampling is performed.
# To run TICPSequential:
# 1. Run the `download_kitti.py` script.
# python3 download_kitti.py
# This will download a city sequence in examples/test_data/open3d_downloads/datasets/kitti_samples/
# if it does not exists, and save the processed frames in `/output/`.
# 2. If you are downloading it anywhere else, change the `dataset_path` accordingly.
# 3. Go to build/bin/examples directory.
# 4. Run the following command:
# ./TICPSequential CPU:0 ../../../examples/cpp/registration_example_util/TICPOdomConfigKitti.txt
# Change CPU:0 to CUDA:0 for running it on primary GPU.
# option to turn ON / OFF visualization:
visualization = ON
visualization_min = -1.5
visualization_max = 1.5
# Verbosity can be Info, Debug.
verbosity = Info
# Path of the downloaded dataset containing frames in .pcd or .ply format.
dataset_path = ../../../examples/test_data/open3d_downloads/datasets/kitti_samples/output/
# Range of frames is start_index to end_index.
start_index = 0
end_index = 1000
# Registration method can be PointToPoint or PointToPlane.
registration_method = PointToPlane
# Multi-Scale ICP parameters:
# Scale 1:
voxel_size = 0.8
search_radii = 1.2
criteria.relative_fitness = 0.01
criteria.relative_rmse = 0.01
criteria.max_iterations = 10
# Scale 2:
voxel_size = 0.5
search_radii = 1.0
criteria.relative_fitness = 0.001
criteria.relative_rmse = 0.001
criteria.max_iterations = 10
# One can also add more scales ...
|
