Hi
I'm trying to make building models using your project.
but cannot get a decent result like your examples
can i get any advice or tip?

this is my input pointcloud building

these are RANSAC parameters ( all default )

before detect Planars, i run Estimate normals with deafult setting

this is the RANSAC result. it looks good to me

After saving the ply and vg files, run make_pc_dataset.py script and reconstruct.py
all parameters in config.yaml are default except normalize and append_bottom

cell_complex.visualize() result

this is final result
All structures on the roof of the building have disappeared.

Hi, @chenzhaiyu ,

Where could I find the Short tutorial as mentioned in readme.md ?

Thanks~

Hi there!

Firstly, I'd like to express my appreciation for the fantastic work on the Points2Poly—it's truly impressive!

Now, onto my main question: I recently attempted to run Points2Poly on a simple structure with a water-tight layout and one internal wall. However, the output file didn't include the internal wall. I'm curious to know if it's possible to reconstruct internal walls and components like stairs using Points2Poly. If so, could you guide me on the necessary modifications in the code and where specifically in the codebase I should make these changes?

here is the simple model that i tried:

1. this is my original model made of simple geometries.

2. here is the the internal wall.

3. here is the point cloud:

4. here is the output with the cut cross section of wall area.

as you can see there is no wall in the output. can you explain this please!

The from sage.all import polytopes, QQ, RR, Polyhedron statement in complex.py is causing an error: ModuleNotFoundError: No module named 'sage.all'. My computer is running Windows 10, and I'm using PyCharm as my development platform.
The version of Sage I installed is only 0.0.0. Is this because of the Windows operating system?

Hi, for the custom dataset, from my point clouds I am able to create the npy files. Would you be able to explain how the folder 05_query_pts should be created? make_pc_dataset.py only creates the .npy and .xyz.npy files. How would I create the files in the folder 05_query_pts? When I run reconstruct.py, it cannot find 05_query_pts folder.
Thank you.

Hi, I'm learning your method but I have some question, can I get any tips?
this is my environment

(1)
I download helsinki_mini datasets and helsinki_fullview_model，and I run reconstructed.py to get six *obj building though it reports some warnings.

reconstruct.log
but I can't get result when I run evaluate.py


(2)
I try to make building models from my point cloud.
I reconstruct a 3D model of a district using UAV and I extract the building by terrasolid. The cavities are caused by trees. this is my input building in mapple.

this is the RANSAC result.(all default)

I follow the step in readme.md to product my datasets and I run reconstruct.py. It seems running but I can't get the ‘reconstructed’ folder.
I only change the dataset name in config

reconstruct2.log

Hi, this is amazing work! I had a question for the custom dataset part, is there an alternative python library/tool to use instead of mapple? I am using windows and the mapple executable for windows seem to be outdated. Is there any other alternative to this especially that is python based? Thank you.

Your research is excellent! I'm interested in your work. when I use your code to reconstruct a real-world building from point cloud,the code can construct cell complex.

[INFO] - cell complex constructed: 0.12 s
[INFO] - number of planes: 6
[INFO] - number of cells: 27

but a ERROR will arise as follows:

[INFO] - cut performed: 0.00 s
[INFO] - cut_value: 0.00
[INFO] - number of extracted cells: 0
[ERROR] - no reachable cells. aborting

Whether this error is caused by the fact that I used your helsinki_fullview model,and the model does not apply to my point clouds?
Do you have any advice for me to solve the problem?

Your work is excellent, and I managed to run through the examples. I would like to ask how to reconstruct my own point cloud file. Now I only have some .PLY files.

Hi,
First of all, thank you very much for providing the code for Points2Poly!
I have been trying to use your work in a docker environment but have systematic crashes when trying to use the reconstruction script on the Helsinki dataset.

Here is the error I get

Traceback (most recent call last):
  File "reconstruct.py", line 51, in reconstruct_full
    cell_complex = create_cell_complex(filepath,
  File "/app/utils.py", line 65, in create_cell_complex
    cell_complex = CellComplex(vertex_group.planes, vertex_group.aabbs, vertex_group.obbs, vertex_group.points_grouped,
  File "/home/appuser/.local/lib/python3.8/site-packages/abspy/complex.py", line 83, in __init__
    self.cells = [self._construct_initial_cell()]  # list of QQ
  File "/home/appuser/.local/lib/python3.8/site-packages/abspy/complex.py", line 105, in _construct_initial_cell
    return polytopes.cube(intervals=[[QQ(self.initial_bound[0][i]), QQ(self.initial_bound[1][i])] for i in range(3)])
TypeError: cube() got an unexpected keyword argument 'intervals'

It seems to be correlated with sagemath, but I am not familiar with this lib, and the documentation still mentions the "intervals" keyword, so I guess the real issue maybe elsewhere. Do you have any hints regarding this error?

Thanks a lot!

For reproducibility, I show the dockerfile and the list of instructions done hereafter.

Dockerfile:

FROM nvidia/cuda:11.3.1-devel-ubuntu20.04

ENV DEBIAN_FRONTEND noninteractive

WORKDIR /app
COPY . /app

RUN apt-get -y update \
    && apt-get install -y software-properties-common \
    && apt-get -y update \
    && add-apt-repository universe
RUN apt-get -y update
RUN apt-get -y install nano git python3 python3-pip wget libgl1-mesa-glx libegl1-mesa libxrandr2 libxrandr2 libxss1 libxcursor1 libxcomposite1 libasound2 libxi6 libxtst6 curl sagemath

RUN adduser -u 5678 --disabled-password --gecos "" appuser && chown -R appuser /app
USER appuser

CMD ["python3", "main.py"]

List of commands once the docker is launched:

pip install ./abspy
pip install -r points2surf/requirements.txt
pip install -r requirements.txt
python download.py dataset_name='helsinki_mini' model_name='helsinki_fullview'
python reconstruct.py dataset_name='helsinki_mini' model_name='helsinki_fullview'

I think your work is very meaningful.How can I run my own .ply file? In the demo of reconstruct.py, the dataset need 04_pts and 06_vertex_group,if I only have a .ply file,if I can get the pts and vertex_group data?

Thank you for this amazing work.

I was wondering if this algorithm works well with indoor planes. I have a indoor dataset captured using rgbd camera. I want to simplify the mesh into planar segments. I previously tried polyfit and was only able to get a watertight planar mesh around the structure of the building rather than indoor planar segments like walls and desks. Could you help me with this? Thank you in advance.

"Hello, I have point cloud data in pcd format. Can I convert the point cloud to ply format and use your reconstruction model to reconstruct the point cloud?"

Hi, thanks a lot for sharing such a brilliant work!
I want to run the demo code and test its function after preparing the dependencies as the instructions said. When I run:
python reconstruct.py dataset_name='helsinki_mini' model_name='helsinki_fullview'
then I meet:

[2023-12-13 21:30:14,735][root][INFO] - processing /home/why/points2poly/datasets/helsinki_mini/06_vertex_group/2.vg
Error executing job with overrides: ['dataset_name=helsinki_mini', 'model_name=helsinki_fullview']
Traceback (most recent call last):
  File "/home/why/points2poly/reconstruct.py", line 51, in reconstruct_full
    cell_complex = create_cell_complex(filepath,
  File "/home/why/points2poly/utils.py", line 69, in create_cell_complex
    cell_complex.refine_planes(theta=theta, epsilon=epsilon)
  File "/home/why/anaconda3/envs/point/lib/python3.10/site-packages/abspy/complex.py", line 127, in refine_planes
    raise ValueError('point coordinates are needed for plane refinement')
ValueError: point coordinates are needed for plane refinement

Set the environment variable HYDRA_FULL_ERROR=1 for a complete stack trace.

I have no idea about this error. Could you please share your thoughts about it?

Regards.

Helsinki dataset cannot be downloaded from OneDrive. Has the link expired?

Thank you for the great work!
What is the minimum performance of the graphics card used in your experiment? And how to handle my own building point cloud data set for training?

Hello,
I would like to run the current approach on a custom dataset. My dataset of buildings contain no bottoms for the buildings. You mentioned in the paper that for the real world Shenzhen dataset you used Helsinki no-bottom model. Can you please provide the trained model?

Also I expect that I will use it the same way as provided in the readme. Just adding the model url and name in the config file, am I correct?

(point2poly) [lilonghui_liujing@gpu01 points2poly]$ python reconstruct.py dataset_name='helsinki_mini' model_name='helsinki_fullview'
/share/home/lilonghui/lilonghui_liujing/zhengxin/environment/anaconda3/envs/point2poly/lib/python3.9/site-packages/sklearn/utils/multiclass.py:14: DeprecationWarning: Please use spmatrix from the scipy.sparse namespace, the scipy.sparse.base namespace is deprecated.
from scipy.sparse.base import spmatrix
/share/home/lilonghui/lilonghui_liujing/zhengxin/environment/anaconda3/envs/point2poly/lib/python3.9/site-packages/sklearn/utils/optimize.py:18: DeprecationWarning: Please use line_search_wolfe2 from the scipy.optimize namespace, the scipy.optimize.linesearch namespace is deprecated.
from scipy.optimize.linesearch import line_search_wolfe2, line_search_wolfe1
/share/home/lilonghui/lilonghui_liujing/zhengxin/environment/anaconda3/envs/point2poly/lib/python3.9/site-packages/sklearn/utils/optimize.py:18: DeprecationWarning: Please use line_search_wolfe1 from the scipy.optimize namespace, the scipy.optimize.linesearch namespace is deprecated.
from scipy.optimize.linesearch import line_search_wolfe2, line_search_wolfe1
[2023-07-04 09:36:53,885][root][INFO] - processing /share/home/lilonghui/lilonghui_liujing/zhengxin/points2poly/datasets/helsinki_mini/06_vertex_group/5.vg
[2023-07-04 09:36:54,723][root][INFO] - refining planar primitives
[2023-07-04 09:36:54,732][root][INFO] - 8 pairs of planes merged
[2023-07-04 09:36:54,732][root][INFO] - prioritising planar primitives
[2023-07-04 09:36:54,733][root][INFO] - constructing cell complex
100%|███████████████████████████████████████████████████████████████████████████████| 23/23 [00:01<00:00, 12.46it/s]
[2023-07-04 09:36:56,581][root][INFO] - cell complex constructed: 1.85 s
[2023-07-04 09:36:56,581][root][INFO] - number of planes: 23
[2023-07-04 09:36:56,581][root][INFO] - number of cells: 129
[2023-07-04 09:36:56,718][root][INFO] - processing /share/home/lilonghui/lilonghui_liujing/zhengxin/points2poly/datasets/helsinki_mini/06_vertex_group/4.vg
[2023-07-04 09:36:56,796][root][INFO] - refining planar primitives
[2023-07-04 09:36:56,804][root][INFO] - 8 pairs of planes merged
[2023-07-04 09:36:56,804][root][INFO] - prioritising planar primitives
[2023-07-04 09:36:56,805][root][INFO] - constructing cell complex
100%|███████████████████████████████████████████████████████████████████████████████| 27/27 [00:03<00:00, 6.80it/s]
[2023-07-04 09:37:00,778][root][INFO] - cell complex constructed: 3.97 s
[2023-07-04 09:37:00,779][root][INFO] - number of planes: 27
[2023-07-04 09:37:00,804][root][INFO] - number of cells: 235
[2023-07-04 09:37:00,960][root][INFO] - processing /share/home/lilonghui/lilonghui_liujing/zhengxin/points2poly/datasets/helsinki_mini/06_vertex_group/6.vg
[2023-07-04 09:37:01,037][root][INFO] - refining planar primitives
[2023-07-04 09:37:01,073][root][INFO] - 34 pairs of planes merged
[2023-07-04 09:37:01,073][root][INFO] - prioritising planar primitives
[2023-07-04 09:37:01,076][root][INFO] - constructing cell complex
100%|███████████████████████████████████████████████████████████████████████████████| 53/53 [00:06<00:00, 8.21it/s]
[2023-07-04 09:37:07,530][root][INFO] - cell complex constructed: 6.45 s
[2023-07-04 09:37:07,530][root][INFO] - number of planes: 53
[2023-07-04 09:37:07,530][root][INFO] - number of cells: 387
[2023-07-04 09:37:07,680][root][INFO] - processing /share/home/lilonghui/lilonghui_liujing/zhengxin/points2poly/datasets/helsinki_mini/06_vertex_group/3.vg
[2023-07-04 09:37:07,756][root][INFO] - refining planar primitives
[2023-07-04 09:37:07,786][root][INFO] - 24 pairs of planes merged
[2023-07-04 09:37:07,786][root][INFO] - prioritising planar primitives
[2023-07-04 09:37:07,788][root][INFO] - constructing cell complex
100%|███████████████████████████████████████████████████████████████████████████████| 31/31 [00:02<00:00, 14.68it/s]
[2023-07-04 09:37:09,901][root][INFO] - cell complex constructed: 2.11 s
[2023-07-04 09:37:09,901][root][INFO] - number of planes: 31
[2023-07-04 09:37:09,901][root][INFO] - number of cells: 116
[2023-07-04 09:37:10,020][root][INFO] - processing /share/home/lilonghui/lilonghui_liujing/zhengxin/points2poly/datasets/helsinki_mini/06_vertex_group/2.vg
[2023-07-04 09:37:10,113][root][INFO] - refining planar primitives
[2023-07-04 09:37:10,144][root][INFO] - 22 pairs of planes merged
[2023-07-04 09:37:10,144][root][INFO] - prioritising planar primitives
[2023-07-04 09:37:10,146][root][INFO] - constructing cell complex
100%|███████████████████████████████████████████████████████████████████████████████| 35/35 [00:02<00:00, 17.20it/s]
[2023-07-04 09:37:12,182][root][INFO] - cell complex constructed: 2.04 s
[2023-07-04 09:37:12,183][root][INFO] - number of planes: 35
[2023-07-04 09:37:12,183][root][INFO] - number of cells: 130
[2023-07-04 09:37:12,263][root][INFO] - processing /share/home/lilonghui/lilonghui_liujing/zhengxin/points2poly/datasets/helsinki_mini/06_vertex_group/1.vg
[2023-07-04 09:37:12,299][root][INFO] - refining planar primitives
[2023-07-04 09:37:12,307][root][INFO] - 12 pairs of planes merged
[2023-07-04 09:37:12,307][root][INFO] - prioritising planar primitives
[2023-07-04 09:37:12,308][root][INFO] - constructing cell complex
100%|███████████████████████████████████████████████████████████████████████████████| 15/15 [00:01<00:00, 14.48it/s]
[2023-07-04 09:37:13,344][root][INFO] - cell complex constructed: 1.04 s
[2023-07-04 09:37:13,344][root][INFO] - number of planes: 15
[2023-07-04 09:37:13,344][root][INFO] - number of cells: 78
evaluating on dataset helsinki_mini
Random Seed: 40938661
getting information for 6 shapes
evaluating 1075 patches
0%| | 0/6 [00:00<?, ?it/s]Warning: imp_surf_dist_ms must be converted to float32
Warning: imp_surf_dist_ms must be converted to float32
Warning: imp_surf_dist_ms must be converted to float32
Warning: imp_surf_dist_ms must be converted to float32
Warning: imp_surf_dist_ms must be converted to float32
Warning: imp_surf_dist_ms must be converted to float32
0%| | 0/6 [00:00<?, ?it/s]
Error executing job with overrides: ['dataset_name=helsinki_mini', 'model_name=helsinki_fullview']
Traceback (most recent call last):
File "/share/home/lilonghui/lilonghui_liujing/zhengxin/points2poly/reconstruct.py", line 71, in reconstruct_full
infer_sdf(cfg)
File "/share/home/lilonghui/lilonghui_liujing/zhengxin/points2poly/utils.py", line 166, in infer_sdf
points_to_surf_eval.points_to_surf_eval(cfg)
File "/share/home/lilonghui/lilonghui_liujing/zhengxin/points2poly/points2surf/source/points_to_surf_eval.py", line 358, in points_to_surf_eval
for batch_data in tqdm(dataloader):
File "/share/home/lilonghui/lilonghui_liujing/zhengxin/environment/anaconda3/envs/point2poly/lib/python3.9/site-packages/tqdm/std.py", line 1176, in iter
for obj in iterable:
File "/share/home/lilonghui/lilonghui_liujing/zhengxin/environment/anaconda3/envs/point2poly/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 633, in next
data = self._next_data()
File "/share/home/lilonghui/lilonghui_liujing/zhengxin/environment/anaconda3/envs/point2poly/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 1345, in _next_data
return self._process_data(data)
File "/share/home/lilonghui/lilonghui_liujing/zhengxin/environment/anaconda3/envs/point2poly/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 1371, in _process_data
data.reraise()
File "/share/home/lilonghui/lilonghui_liujing/zhengxin/environment/anaconda3/envs/point2poly/lib/python3.9/site-packages/torch/_utils.py", line 644, in reraise
raise exception
TypeError: Caught TypeError in DataLoader worker process 0.
Original Traceback (most recent call last):
File "/share/home/lilonghui/lilonghui_liujing/zhengxin/environment/anaconda3/envs/point2poly/lib/python3.9/site-packages/torch/utils/data/_utils/worker.py", line 308, in _worker_loop
data = fetcher.fetch(index)
File "/share/home/lilonghui/lilonghui_liujing/zhengxin/environment/anaconda3/envs/point2poly/lib/python3.9/site-packages/torch/utils/data/_utils/fetch.py", line 51, in fetch
data = [self.dataset[idx] for idx in possibly_batched_index]
File "/share/home/lilonghui/lilonghui_liujing/zhengxin/environment/anaconda3/envs/point2poly/lib/python3.9/site-packages/torch/utils/data/_utils/fetch.py", line 51, in
data = [self.dataset[idx] for idx in possibly_batched_index]
File "/share/home/lilonghui/lilonghui_liujing/zhengxin/points2poly/points2surf/source/data_loader.py", line 359, in getitem
get_patch_points(shape=shape, query_point=imp_surf_query_point_ms)
File "/share/home/lilonghui/lilonghui_liujing/zhengxin/points2poly/points2surf/source/data_loader.py", line 335, in get_patch_points
patch_pts_ids = point_cloud.get_patch_kdtree(
File "/share/home/lilonghui/lilonghui_liujing/zhengxin/points2poly/points2surf/source/base/point_cloud.py", line 174, in get_patch_kdtree
pts_dists_ms, patch_pts_ids = kdtree.query(x=query_point, k=points_per_patch, n_jobs=n_jobs)
File "_ckdtree.pyx", line 786, in scipy.spatial._ckdtree.cKDTree.query
File "_ckdtree.pyx", line 388, in scipy.spatial._ckdtree.get_num_workers
TypeError: Unexpected keyword argument {'n_jobs': 1}

Set the environment variable HYDRA_FULL_ERROR=1 for a complete stack trace.

Hello, this is a great job. However, I have encountered a problem while reconstructing my own data, "Extract planar primitives from point clouds with Mapple." When I need to reconstruct a lot of data on my own, manually obtaining planar primitives will take a lot of time in this step. Is there any other way to complete this task instead of doing it manually? Looking forward to your answer.

Describe the bug
A clear and concise description of what the bug is.

To Reproduce
Steps to reproduce the behavior:

1. Input '...'
2. Run '....'
3. See error

Expected behavior
A clear and concise description of what you expected to happen.

Screenshots
If applicable, add screenshots to help explain your problem.

Environment
Describe the environment (e.g., package version) if pertinent.

Additional context
Add any other context about the problem here.

Do I need to install sage? sage looks so big.

(myria3d) root@e7c988bbfef7:/data/points2poly# python reconstruct.py dataset_name='helsinki_mini' model_name='helsinki_fullview'
Traceback (most recent call last):
File "/data/points2poly/reconstruct.py", line 29, in
from utils import create_cell_complex, create_query_points, extract_surface, infer_sdf
File "/data/points2poly/utils.py", line 14, in
from abspy import VertexGroup, CellComplex, AdjacencyGraph
File "/root/miniconda3/envs/myria3d/lib/python3.9/site-packages/abspy/init.py", line 5, in
from .complex import *
File "/root/miniconda3/envs/myria3d/lib/python3.9/site-packages/abspy/complex.py", line 28, in
from sage.all import polytopes, QQ, RR, Polyhedron
ModuleNotFoundError: No module named 'sage'
(myria3d) root@e7c988bbfef7:/data/points2poly#

Hello ,It's truly a great job! Would you please share the datasets of shenzhen on this paper! Looking forward to your help！

I'm getting this overflow warning/error and it's also happening in helsinki_mini example:
points2poly/utils.py:32: RuntimeWarning: overflow encountered in exp
return 1 / (1 + np.exp(-x))