Datasets:

handpose_annotations / HOT3DHUGGFACE /data_loaders /ObjectPose3dProvider.py
FKZZddd's picture
Upload folder using huggingface_hub
dbf8921 verified
Raw
History Blame Contribute Delete
6.48 kB
# Copyright (c) Meta Platforms, Inc. and affiliates.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import csv
from dataclasses import dataclass
from typing import Any, Dict, List, Optional, Set
import numpy as np
from projectaria_tools.core.sensor_data import TimeDomain, TimeQueryOptions # @manual
from projectaria_tools.core.sophus import SE3 # @manual
from .constants import POSE_DATA_CSV_COLUMNS
from .loader_poses_utils import check_csv_columns
from .pose_utils import lookup_timestamp
@dataclass
class ObjectPose3d:
"""
Class to store pose of a single entity (object/headset/hand)
"""
T_world_object: Optional[SE3] = None
@dataclass
class ObjectPose3dCollection:
"""
Class to store the poses for a given timestamp
"""
timestamp_ns: int
poses: Dict[str, ObjectPose3d]
@property
def object_uid_list(self) -> Set[str]:
return set(self.poses.keys())
ObjectPose3dTrajectory = Dict[int, ObjectPose3dCollection]
@dataclass
class ObjectPose3dCollectionWithDt:
pose3d_collection: ObjectPose3dCollection
time_delta_ns: int
class ObjectPose3dProvider(object):
def __init__(self, pose3d_trajectory: ObjectPose3dTrajectory):
self._pose3d_trajectory: ObjectPose3dTrajectory = pose3d_trajectory
#排序时间戳
self._sorted_timestamp_ns_list: List[int] = sorted(
self._pose3d_trajectory.keys()
)
#所有出现过的物体ID
self._object_uids_with_poses: Set = {
x for v in self._pose3d_trajectory.values() for x in v.object_uid_list
}
#像属性一样访问,不用加括号
@property
def timestamp_ns_list(self) -> List[int]:
return self._sorted_timestamp_ns_list
@property
def object_uids_with_poses(self) -> Set[str]:
return set(self._object_uids_with_poses)
def get_data_statistics(self) -> Dict[str, Any]:
"""
Returns the stats of the trajectory
"""
stats = {}
stats["num_frames"] = len(self._sorted_timestamp_ns_list)
stats["num_objects"] = len(self._object_uids_with_poses)
stats["object_uids"] = [str(x) for x in self._object_uids_with_poses]
return stats
def get_pose_at_timestamp(
self,
timestamp_ns: int,
time_query_options: TimeQueryOptions,
time_domain: TimeDomain,
acceptable_time_delta: Optional[int] = None,
) -> Optional[ObjectPose3dCollectionWithDt]:
"""
Return the list of poses available at the given timestamp
"""
if time_domain is not TimeDomain.TIME_CODE:
raise ValueError("Value other than TimeDomain.TIME_CODE not yet supported.")
pose3d_collection, time_delta_ns = lookup_timestamp(
time_indexed_dict=self._pose3d_trajectory,
sorted_timestamp_list=self._sorted_timestamp_ns_list,
query_timestamp=timestamp_ns,
time_query_options=time_query_options,
)
if (
pose3d_collection is None
or time_delta_ns is None
or (
acceptable_time_delta is not None
and abs(time_delta_ns) > acceptable_time_delta
)
):
return None
else:
return ObjectPose3dCollectionWithDt(
pose3d_collection=pose3d_collection, time_delta_ns=time_delta_ns
)
#生成了一个pose3d_trajectory里面有物品在每个时间下的位置
def load_object_pose_trajectory_from_csv(filename: str) -> ObjectPose3dTrajectory:
"""Load Dynamic Objects meta data from a CSV file.
Keyword arguments:
filename -- the csv file i.e. sequence_folder + "/dynamic_objects.csv"
"""
#空字典,后面逐行读 CSV 往里填数据
pose3d_trajectory: ObjectPose3dTrajectory = {}
# Open the CSV file for reading
with open(filename, "r") as f:
reader = csv.reader(f)
# Read the header row
#读表头
header = next(reader)
# Ensure we have the desired columns
check_csv_columns(header, POSE_DATA_CSV_COLUMNS)
# Read the rest of the rows in the CSV file
for row in reader:
#位置
translation = [
#"t_wo_x[m]" 这个字符串在 header 列表的第几位? 然后读取
row[header.index("t_wo_x[m]")],
row[header.index("t_wo_y[m]")],
row[header.index("t_wo_z[m]")],
]
#朝向
quaternion_xyz = [
##"q_wo_x 这个字符串在 header 列表的第几位? 然后读取
row[header.index("q_wo_x")],
row[header.index("q_wo_y")],
row[header.index("q_wo_z")],
]
quaternion_w = row[header.index("q_wo_w")]
timestamp_ns = int(row[header.index("timestamp[ns]")])
object_uid = str(row[header.index("object_uid")])
#把位置和朝向转换成 SE3 变换矩阵,se3 物品位置和朝向的信息
T_world_object = SE3.from_quat_and_translation(
float(quaternion_w),
np.array([float(o) for o in quaternion_xyz]),
np.array([float(o) for o in translation]),
)[0]
pose3d = ObjectPose3d(T_world_object=T_world_object)
#确保timestamp_ns有位置
if timestamp_ns not in pose3d_trajectory:
pose3d_trajectory[timestamp_ns] = ObjectPose3dCollection(
timestamp_ns=timestamp_ns, poses={}
)
pose3d_trajectory[timestamp_ns].poses[object_uid] = pose3d
return pose3d_trajectory
def load_pose_provider_from_csv(filename: str) -> ObjectPose3dProvider:
"""
Load the ObjectPose3dProvider from a csv file
"""
return ObjectPose3dProvider(
pose3d_trajectory=load_object_pose_trajectory_from_csv(filename)
)