mirror of
https://github.com/nerfstudio-project/nerfstudio.git
synced 2026-11-20 14:45:38 +08:00
148 lines
5.6 KiB
Python
148 lines
5.6 KiB
Python
# Copyright 2022 The Nerfstudio Team. All rights reserved.
|
|
#
|
|
# 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.
|
|
|
|
"""
|
|
Pose and Intrinsics Optimizers
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
import functools
|
|
from dataclasses import dataclass, field
|
|
from typing import Type, Union
|
|
|
|
import torch
|
|
import tyro
|
|
from torch import nn
|
|
from torchtyping import TensorType
|
|
from typing_extensions import Literal, assert_never
|
|
|
|
from nerfstudio.cameras.lie_groups import exp_map_SE3, exp_map_SO3xR3
|
|
from nerfstudio.configs import base_config as cfg
|
|
from nerfstudio.engine.optimizers import AdamOptimizerConfig
|
|
from nerfstudio.engine.schedulers import SchedulerConfig
|
|
from nerfstudio.utils import poses as pose_utils
|
|
|
|
|
|
@dataclass
|
|
class CameraOptimizerConfig(cfg.InstantiateConfig):
|
|
"""Configuration of optimization for camera poses."""
|
|
|
|
_target: Type = field(default_factory=lambda: CameraOptimizer)
|
|
|
|
mode: Literal["off", "SO3xR3", "SE3"] = "off"
|
|
"""Pose optimization strategy to use. If enabled, we recommend SO3xR3."""
|
|
|
|
position_noise_std: float = 0.0
|
|
"""Noise to add to initial positions. Useful for debugging."""
|
|
|
|
orientation_noise_std: float = 0.0
|
|
"""Noise to add to initial orientations. Useful for debugging."""
|
|
|
|
optimizer: AdamOptimizerConfig = AdamOptimizerConfig(lr=6e-4, eps=1e-15)
|
|
"""ADAM parameters for camera optimization."""
|
|
|
|
scheduler: SchedulerConfig = SchedulerConfig(max_steps=10000)
|
|
"""Learning rate scheduler for camera optimizer.."""
|
|
|
|
param_group: tyro.conf.Suppress[str] = "camera_opt"
|
|
"""Name of the parameter group used for pose optimization. Can be any string that doesn't conflict with other
|
|
groups."""
|
|
|
|
|
|
class CameraOptimizer(nn.Module):
|
|
"""Layer that modifies camera poses to be optimized as well as the field during training."""
|
|
|
|
config: CameraOptimizerConfig
|
|
|
|
def __init__(
|
|
self,
|
|
config: CameraOptimizerConfig,
|
|
num_cameras: int,
|
|
device: Union[torch.device, str],
|
|
**kwargs, # pylint: disable=unused-argument
|
|
) -> None:
|
|
super().__init__()
|
|
self.config = config
|
|
self.num_cameras = num_cameras
|
|
self.device = device
|
|
|
|
# Initialize learnable parameters.
|
|
if self.config.mode == "off":
|
|
pass
|
|
elif self.config.mode in ("SO3xR3", "SE3"):
|
|
self.pose_adjustment = torch.nn.Parameter(torch.zeros((num_cameras, 6), device=device))
|
|
self.blur_adjustment = torch.nn.Parameter(torch.zeros((num_cameras, 6), device=device))
|
|
else:
|
|
assert_never(self.config.mode)
|
|
|
|
# Initialize pose noise; useful for debugging.
|
|
if config.position_noise_std != 0.0 or config.orientation_noise_std != 0.0:
|
|
assert config.position_noise_std >= 0.0 and config.orientation_noise_std >= 0.0
|
|
std_vector = torch.tensor(
|
|
[config.position_noise_std] * 3 + [config.orientation_noise_std] * 3, device=device
|
|
)
|
|
self.pose_noise = exp_map_SE3(torch.normal(torch.zeros((num_cameras, 6), device=device), std_vector))
|
|
else:
|
|
self.pose_noise = None
|
|
|
|
def sample_blur_correction(self, indices: TensorType["num_cameras"]):
|
|
"""
|
|
Samples the blur parameters and returns transformation matrices for the origins and directions
|
|
"""
|
|
blur_adjs = self.blur_adjustment[indices, :] * torch.rand((indices.shape[0], 1), device=self.device)
|
|
# print(
|
|
# "trans,rot norm",
|
|
# torch.mean(torch.linalg.norm(blur_adjs[:, :3], dim=0)),
|
|
# torch.mean(torch.linalg.norm(blur_adjs[:, 3:], dim=0)),
|
|
# )
|
|
return exp_map_SO3xR3(blur_adjs)
|
|
# origin_deltas = blur_adjs[:, :3] * torch.rand((blur_adjs.shape[0], 1), device=self.device)
|
|
# direction_amnts = torch.linalg.norm(blur_adjs[:, 3:])
|
|
|
|
return origin_deltas, None
|
|
|
|
def forward(
|
|
self,
|
|
indices: TensorType["num_cameras"],
|
|
) -> TensorType["num_cameras", 3, 4]:
|
|
"""Indexing into camera adjustments.
|
|
Args:
|
|
indices: indices of Cameras to optimize.
|
|
Returns:
|
|
Tranformation matrices from optimized camera coordinates coordinates
|
|
to given camera coordinates.
|
|
"""
|
|
outputs = []
|
|
|
|
# Apply learned transformation delta.
|
|
if self.config.mode == "off":
|
|
pass
|
|
elif self.config.mode == "SO3xR3":
|
|
outputs.append(exp_map_SO3xR3(self.pose_adjustment[indices, :]))
|
|
elif self.config.mode == "SE3":
|
|
outputs.append(exp_map_SE3(self.pose_adjustment[indices, :]))
|
|
else:
|
|
assert_never(self.config.mode)
|
|
|
|
# Apply initial pose noise.
|
|
if self.pose_noise is not None:
|
|
outputs.append(self.pose_noise[indices, :, :])
|
|
|
|
# Return: identity if no transforms are needed, otherwise multiply transforms together.
|
|
if len(outputs) == 0:
|
|
# Note that using repeat() instead of tile() here would result in unnecessary copies.
|
|
return torch.eye(4, device=self.device)[None, :3, :4].tile(indices.shape[0], 1, 1)
|
|
return functools.reduce(pose_utils.multiply, outputs)
|