Skip to content

Commit

Permalink
Merge branch '867-cpp-filtering' into 'master'
Browse files Browse the repository at this point in the history 
867 cars-filter integration

Closes #867

See merge request 3d/cars-park/cars!758
  • Loading branch information
@dyoussef
dyoussef committed Nov 21, 2024
2 parents ef49962 + 0c79e4a commit b42cd21
Show file tree
Hide file tree
Showing 22 changed files with 1,321 additions and 559 deletions.
2 changes: 1 addition & 1 deletion cars/applications/__init__.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -31,7 +31,7 @@
from . import holes_detection # noqa: F401
from . import point_cloud_denoising # noqa: F401
from . import point_cloud_fusion # noqa: F401
from . import point_cloud_outliers_removing # noqa: F401
from . import point_cloud_outlier_removal # noqa: F401
from . import rasterization # noqa: F401
from . import resampling # noqa: F401
from . import sparse_matching # noqa: F401
Expand Down
4 changes: 2 additions & 2 deletions ...point_cloud_outliers_removing/__init__.py → ...s/point_cloud_outlier_removal/__init__.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -23,8 +23,8 @@
"""
# flake8: noqa: F401

from cars.applications.point_cloud_outliers_removing.pc_out_removing import (
PointCloudOutliersRemoving,
from cars.applications.point_cloud_outlier_removal.pc_out_removal import (
PointCloudOutlierRemoval,
)

from . import small_components, statistical
252 changes: 106 additions & 146 deletions ...tliers_removing/outlier_removing_tools.py → ..._outlier_removal/outlier_removal_tools.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -19,26 +19,28 @@
# limitations under the License.
#
"""
This module contains functions used in outlier removing
This module contains functions used in outlier removal
"""

# Standard imports
import logging
from typing import List, Tuple, Union

# Third party imports
import numpy as np
import outlier_filter # pylint:disable=E0401
import pandas
import xarray as xr
from scipy.spatial import cKDTree # pylint: disable=no-name-in-module

from cars.applications.point_cloud_fusion.point_cloud_tools import filter_cloud

# CARS imports
from cars.core import constants as cst
from cars.core import projection

# ##### Small components filtering ######
# ##### Small component filtering ######


def small_components_filtering(
def small_component_filtering(
cloud: pandas.DataFrame,
connection_val: float,
nb_pts_threshold: int,
Expand All @@ -64,9 +66,20 @@ def small_components_filtering(
:return: Tuple made of the filtered cloud and
the removed elements positions in their epipolar images
"""
cloud_xyz = cloud.loc[:, [cst.X, cst.Y, cst.Z]].values
index_elt_to_remove = detect_small_components(
cloud_xyz, connection_val, nb_pts_threshold, clusters_distance_threshold

clusters_distance_threshold_float = (
np.nan
if clusters_distance_threshold is None
else clusters_distance_threshold
)

index_elt_to_remove = outlier_filter.pc_small_component_outlier_filtering(
cloud.loc[:, cst.X].values,
cloud.loc[:, cst.Y].values,
cloud.loc[:, cst.Z].values,
radius=connection_val,
min_cluster_size=nb_pts_threshold,
clusters_distance_threshold=clusters_distance_threshold_float,
)

return filter_cloud(cloud, index_elt_to_remove, filtered_elt_pos)
Expand Down Expand Up @@ -175,7 +188,7 @@ def detect_small_components(
# ##### statistical filtering ######


def statistical_outliers_filtering(
def statistical_outlier_filtering(
cloud: pandas.DataFrame,
k: int,
dev_factor: float,
Expand All @@ -198,9 +211,14 @@ def statistical_outliers_filtering(
:return: Tuple made of the filtered cloud and
the removed elements positions in their epipolar images
"""
cloud_xyz = cloud.loc[:, [cst.X, cst.Y, cst.Z]].values
index_elt_to_remove = detect_statistical_outliers(
cloud_xyz, k, dev_factor, use_median

index_elt_to_remove = outlier_filter.pc_statistical_outlier_filtering(
cloud.loc[:, cst.X].values,
cloud.loc[:, cst.Y].values,
cloud.loc[:, cst.Z].values,
dev_factor=dev_factor,
k=k,
use_median=use_median,
)

return filter_cloud(cloud, index_elt_to_remove, filtered_elt_pos)
Expand Down Expand Up @@ -247,6 +265,7 @@ def detect_statistical_outliers(
iqr_distances = np.percentile(
mean_neighbors_distances, 75
) - np.percentile(mean_neighbors_distances, 25)

# compute distance threshold and
# apply it to determine which points will be removed
dist_thresh = median_distances + dev_factor * iqr_distances
Expand All @@ -269,151 +288,92 @@ def detect_statistical_outliers(
return detected_points


# ##### common filtering tools ######

def epipolar_small_components(
cloud,
epsg,
min_cluster_size=15,
radius=1.0,
half_window_size=5,
clusters_distance_threshold=np.nan,
):
"""
Filter outliers using the small components method in epipolar geometry
:param epipolar_ds: epipolar dataset to filter
:type epipolar_ds: xr.Dataset
:param epsg: epsg code of the CRS used to compute distances
:type epsg: int
:param statistical_k: k
:type statistical_k: int
:param std_dev_factor: std factor
:type std_dev_factor: float
:param half_window_size: use median and quartile instead of mean and std
:type half_window_size: int
:param use_median: use median and quartile instead of mean and std
:type use_median: bool
:return: filtered dataset
:rtype: xr.Dataset
def filter_cloud(
cloud: pandas.DataFrame,
index_elt_to_remove: List[int],
filtered_elt_pos: bool = False,
) -> Tuple[pandas.DataFrame, Union[None, pandas.DataFrame]]:
"""
Filter all points of the cloud DataFrame
which index is in the index_elt_to_remove list.

If filtered_elt_pos is set to True, the information of the removed elements
positions in their original epipolar images are returned.
projection.points_cloud_conversion_dataset(cloud, epsg)

To do so the cloud DataFrame has to be build
with the 'with_coords' option activated.
if clusters_distance_threshold is None:
clusters_distance_threshold = np.nan

:param cloud: combined cloud
as returned by the create_combined_cloud function
:param index_elt_to_remove: indexes of lines
to filter in the cloud DataFrame
:param filtered_elt_pos: if filtered_elt_pos is set to True,
the removed points positions in their original epipolar images are
returned, otherwise it is set to None
:return: Tuple composed of the filtered cloud DataFrame and
the filtered elements epipolar position information
(or None for the latter if filtered_elt_pos is set to False
or if the cloud Dataframe has not been build with with_coords option)
"""
if filtered_elt_pos and not (
cst.POINTS_CLOUD_COORD_EPI_GEOM_I in cloud.columns
and cst.POINTS_CLOUD_COORD_EPI_GEOM_J in cloud.columns
and cst.POINTS_CLOUD_ID_IM_EPI in cloud.columns
):
logging.warning(
"In filter_cloud: the filtered_elt_pos has been activated but "
"the cloud Datafram has not been build with option with_coords. "
"The positions cannot be retrieved."
)
filtered_elt_pos = False

# retrieve removed points position in their original epipolar images
if filtered_elt_pos:
labels = [
cst.POINTS_CLOUD_COORD_EPI_GEOM_I,
cst.POINTS_CLOUD_COORD_EPI_GEOM_J,
cst.POINTS_CLOUD_ID_IM_EPI,
]

removed_elt_pos_infos = cloud.loc[
cloud.index.values[index_elt_to_remove], labels
].values

removed_elt_pos_infos = pandas.DataFrame(
removed_elt_pos_infos, columns=labels
)
else:
removed_elt_pos_infos = None

# remove points from the cloud
cloud = cloud.drop(index=cloud.index.values[index_elt_to_remove])
outlier_filter.epipolar_small_component_outlier_filtering(
cloud[cst.X],
cloud[cst.Y],
cloud[cst.Z],
min_cluster_size,
radius,
half_window_size,
clusters_distance_threshold,
)

return cloud, removed_elt_pos_infos
return cloud


def add_cloud_filtering_msk(
clouds_list: List[xr.Dataset],
elt_pos_infos: pandas.DataFrame,
mask_label: str,
mask_value: int = 255,
def epipolar_statistical_filtering(
epipolar_ds,
epsg,
k=15,
dev_factor=1.0,
half_window_size=5,
use_median=False,
):
"""
Add a uint16 mask labeled 'mask_label' to the clouds in clouds_list.
(in-line function)
:param clouds_list: Input list of clouds
:param elt_pos_infos: pandas dataframe
composed of cst.POINTS_CLOUD_COORD_EPI_GEOM_I,
cst.POINTS_CLOUD_COORD_EPI_GEOM_J, cst.POINTS_CLOUD_ID_IM_EPI columns
as computed in the create_combined_cloud function.
Those information are used to retrieve the point position
in its original epipolar image.
:param mask_label: label to give to the mask in the datasets
:param mask_value: filtered elements value in the mask
"""

# Verify that the elt_pos_infos is consistent
if (
elt_pos_infos is None
or cst.POINTS_CLOUD_COORD_EPI_GEOM_I not in elt_pos_infos.columns
or cst.POINTS_CLOUD_COORD_EPI_GEOM_J not in elt_pos_infos.columns
or cst.POINTS_CLOUD_ID_IM_EPI not in elt_pos_infos.columns
):
logging.warning(
"Cannot generate filtered elements mask, "
"no information about the point's"
" original position in the epipolar image is given"
)

else:
elt_index = elt_pos_infos.loc[:, cst.POINTS_CLOUD_ID_IM_EPI].to_numpy()

min_elt_index = np.min(elt_index)
max_elt_index = np.max(elt_index)
Filter outliers using the statistical method in epipolar geometry
:param epipolar_ds: epipolar dataset to filter
:type epipolar_ds: xr.Dataset
:param epsg: epsg code of the CRS used to compute distances
:type epsg: int
:param statistical_k: k
:type statistical_k: int
:param std_dev_factor: std factor
:type std_dev_factor: float
:param half_window_size: use median and quartile instead of mean and std
:type half_window_size: int
:param use_median: use median and quartile instead of mean and std
:type use_median: bool
:return: filtered dataset
:rtype: xr.Dataset
if min_elt_index < 0 or max_elt_index > len(clouds_list) - 1:
raise RuntimeError(
"Index indicated in the elt_pos_infos pandas. "
"DataFrame is not coherent with the clouds list given in input"
)

# create and add mask to each element of clouds_list
for cloud_idx, cloud_item in enumerate(clouds_list):
if mask_label not in cloud_item:
nb_row = cloud_item.coords[cst.ROW].data.shape[0]
nb_col = cloud_item.coords[cst.COL].data.shape[0]
msk = np.zeros((nb_row, nb_col), dtype=np.uint16)
else:
msk = cloud_item[mask_label].values
"""

cur_elt_index = np.argwhere(elt_index == cloud_idx)
projection.points_cloud_conversion_dataset(epipolar_ds, epsg)

for elt_pos in range(cur_elt_index.shape[0]):
i = int(
elt_pos_infos.loc[
cur_elt_index[elt_pos],
cst.POINTS_CLOUD_COORD_EPI_GEOM_I,
].iat[0]
)
j = int(
elt_pos_infos.loc[
cur_elt_index[elt_pos],
cst.POINTS_CLOUD_COORD_EPI_GEOM_J,
].iat[0]
)
outlier_filter.epipolar_statistical_outlier_filtering(
epipolar_ds[cst.X],
epipolar_ds[cst.Y],
epipolar_ds[cst.Z],
k,
half_window_size,
dev_factor,
use_median,
)

try:
msk[i, j] = mask_value
except Exception as index_error:
raise RuntimeError(
"Point at location ({},{}) is not accessible "
"in an image of size ({},{})".format(
i, j, msk.shape[0], msk.shape[1]
)
) from index_error

cloud_item[mask_label] = ([cst.ROW, cst.COL], msk)
return epipolar_ds
Loading

0 comments on commit b42cd21

Please sign in to comment.