transforms_3d.py 55.1 KB
Newer Older
dingchang's avatar
dingchang committed
1
# Copyright (c) OpenMMLab. All rights reserved.
zhangwenwei's avatar
zhangwenwei committed
2
import numpy as np
3
import warnings
4
from mmcv import is_tuple_of
5
from mmcv.utils import build_from_cfg
zhangwenwei's avatar
zhangwenwei committed
6

7
from mmdet3d.core import VoxelGenerator
8
9
from mmdet3d.core.bbox import (CameraInstance3DBoxes, DepthInstance3DBoxes,
                               LiDARInstance3DBoxes, box_np_ops)
10
from mmdet.datasets.builder import PIPELINES
zhangwenwei's avatar
zhangwenwei committed
11
from mmdet.datasets.pipelines import RandomFlip
12
from ..builder import OBJECTSAMPLERS
zhangwenwei's avatar
zhangwenwei committed
13
14
15
from .data_augment_utils import noise_per_object_v3_


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
@PIPELINES.register_module()
class RandomDropPointsColor(object):
    r"""Randomly set the color of points to all zeros.

    Once this transform is executed, all the points' color will be dropped.
    Refer to `PAConv <https://github.com/CVMI-Lab/PAConv/blob/main/scene_seg/
    util/transform.py#L223>`_ for more details.

    Args:
        drop_ratio (float): The probability of dropping point colors.
            Defaults to 0.2.
    """

    def __init__(self, drop_ratio=0.2):
        assert isinstance(drop_ratio, (int, float)) and 0 <= drop_ratio <= 1, \
            f'invalid drop_ratio value {drop_ratio}'
        self.drop_ratio = drop_ratio

    def __call__(self, input_dict):
        """Call function to drop point colors.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
            dict: Results after color dropping, \
                'points' key is updated in the result dict.
        """
        points = input_dict['points']
        assert points.attribute_dims is not None and \
            'color' in points.attribute_dims, \
            'Expect points have color attribute'

49
50
51
52
53
54
55
        # this if-expression is a bit strange
        # `RandomDropPointsColor` is used in training 3D segmentor PAConv
        # we discovered in our experiments that, using
        # `if np.random.rand() > 1.0 - self.drop_ratio` consistently leads to
        # better results than using `if np.random.rand() < self.drop_ratio`
        # so we keep this hack in our codebase
        if np.random.rand() > 1.0 - self.drop_ratio:
56
57
58
59
60
61
62
63
64
65
            points.color = points.color * 0.0
        return input_dict

    def __repr__(self):
        """str: Return a string that describes the module."""
        repr_str = self.__class__.__name__
        repr_str += f'(drop_ratio={self.drop_ratio})'
        return repr_str


66
@PIPELINES.register_module()
zhangwenwei's avatar
zhangwenwei committed
67
68
69
70
71
72
73
74
class RandomFlip3D(RandomFlip):
    """Flip the points & bbox.

    If the input dict contains the key "flip", then the flag will be used,
    otherwise it will be randomly decided by a ratio specified in the init
    method.

    Args:
zhangwenwei's avatar
zhangwenwei committed
75
76
77
        sync_2d (bool, optional): Whether to apply flip according to the 2D
            images. If True, it will apply the same flip as that to 2D images.
            If False, it will decide whether to flip randomly and independently
liyinhao's avatar
liyinhao committed
78
            to that of 2D images. Defaults to True.
wuyuefeng's avatar
wuyuefeng committed
79
        flip_ratio_bev_horizontal (float, optional): The flipping probability
liyinhao's avatar
liyinhao committed
80
            in horizontal direction. Defaults to 0.0.
wuyuefeng's avatar
wuyuefeng committed
81
        flip_ratio_bev_vertical (float, optional): The flipping probability
liyinhao's avatar
liyinhao committed
82
            in vertical direction. Defaults to 0.0.
zhangwenwei's avatar
zhangwenwei committed
83
84
    """

wuyuefeng's avatar
wuyuefeng committed
85
86
87
88
89
90
91
    def __init__(self,
                 sync_2d=True,
                 flip_ratio_bev_horizontal=0.0,
                 flip_ratio_bev_vertical=0.0,
                 **kwargs):
        super(RandomFlip3D, self).__init__(
            flip_ratio=flip_ratio_bev_horizontal, **kwargs)
zhangwenwei's avatar
zhangwenwei committed
92
        self.sync_2d = sync_2d
wuyuefeng's avatar
wuyuefeng committed
93
94
95
96
97
98
99
100
101
102
103
        self.flip_ratio_bev_vertical = flip_ratio_bev_vertical
        if flip_ratio_bev_horizontal is not None:
            assert isinstance(
                flip_ratio_bev_horizontal,
                (int, float)) and 0 <= flip_ratio_bev_horizontal <= 1
        if flip_ratio_bev_vertical is not None:
            assert isinstance(
                flip_ratio_bev_vertical,
                (int, float)) and 0 <= flip_ratio_bev_vertical <= 1

    def random_flip_data_3d(self, input_dict, direction='horizontal'):
104
105
106
107
108
109
110
111
112
113
        """Flip 3D data randomly.

        Args:
            input_dict (dict): Result dict from loading pipeline.
            direction (str): Flip direction. Default: horizontal.

        Returns:
            dict: Flipped results, 'points', 'bbox3d_fields' keys are \
                updated in the result dict.
        """
wuyuefeng's avatar
wuyuefeng committed
114
        assert direction in ['horizontal', 'vertical']
115
116
117
118
        # for semantic segmentation task, only points will be flipped.
        if 'bbox3d_fields' not in input_dict:
            input_dict['points'].flip(direction)
            return
119
120
121
122
123
        if len(input_dict['bbox3d_fields']) == 0:  # test mode
            input_dict['bbox3d_fields'].append('empty_box3d')
            input_dict['empty_box3d'] = input_dict['box_type_3d'](
                np.array([], dtype=np.float32))
        assert len(input_dict['bbox3d_fields']) == 1
zhangwenwei's avatar
zhangwenwei committed
124
        for key in input_dict['bbox3d_fields']:
125
126
127
128
129
130
131
132
            if 'points' in input_dict:
                input_dict['points'] = input_dict[key].flip(
                    direction, points=input_dict['points'])
            else:
                input_dict[key].flip(direction)
        if 'centers2d' in input_dict:
            assert self.sync_2d is True and direction == 'horizontal', \
                'Only support sync_2d=True and horizontal flip with images'
133
            w = input_dict['ori_shape'][1]
134
135
            input_dict['centers2d'][..., 0] = \
                w - input_dict['centers2d'][..., 0]
136
137
            # need to modify the horizontal position of camera center
            # along u-axis in the image (flip like centers2d)
138
            # ['cam2img'][0][2] = c_u
139
140
            # see more details and examples at
            # https://github.com/open-mmlab/mmdetection3d/pull/744
141
            input_dict['cam2img'][0][2] = w - input_dict['cam2img'][0][2]
zhangwenwei's avatar
zhangwenwei committed
142
143

    def __call__(self, input_dict):
144
145
146
147
148
149
150
151
152
153
154
        """Call function to flip points, values in the ``bbox3d_fields`` and \
        also flip 2D image and its annotations.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
            dict: Flipped results, 'flip', 'flip_direction', \
                'pcd_horizontal_flip' and 'pcd_vertical_flip' keys are added \
                into result dict.
        """
155
        # flip 2D image and its annotations
zhangwenwei's avatar
zhangwenwei committed
156
        super(RandomFlip3D, self).__call__(input_dict)
zhangwenwei's avatar
zhangwenwei committed
157

zhangwenwei's avatar
zhangwenwei committed
158
        if self.sync_2d:
wuyuefeng's avatar
wuyuefeng committed
159
160
            input_dict['pcd_horizontal_flip'] = input_dict['flip']
            input_dict['pcd_vertical_flip'] = False
zhangwenwei's avatar
zhangwenwei committed
161
        else:
wuyuefeng's avatar
wuyuefeng committed
162
163
164
165
166
167
168
169
170
            if 'pcd_horizontal_flip' not in input_dict:
                flip_horizontal = True if np.random.rand(
                ) < self.flip_ratio else False
                input_dict['pcd_horizontal_flip'] = flip_horizontal
            if 'pcd_vertical_flip' not in input_dict:
                flip_vertical = True if np.random.rand(
                ) < self.flip_ratio_bev_vertical else False
                input_dict['pcd_vertical_flip'] = flip_vertical

171
172
173
        if 'transformation_3d_flow' not in input_dict:
            input_dict['transformation_3d_flow'] = []

wuyuefeng's avatar
wuyuefeng committed
174
175
        if input_dict['pcd_horizontal_flip']:
            self.random_flip_data_3d(input_dict, 'horizontal')
176
            input_dict['transformation_3d_flow'].extend(['HF'])
wuyuefeng's avatar
wuyuefeng committed
177
178
        if input_dict['pcd_vertical_flip']:
            self.random_flip_data_3d(input_dict, 'vertical')
179
            input_dict['transformation_3d_flow'].extend(['VF'])
zhangwenwei's avatar
zhangwenwei committed
180
181
        return input_dict

zhangwenwei's avatar
zhangwenwei committed
182
    def __repr__(self):
183
        """str: Return a string that describes the module."""
wuyuefeng's avatar
wuyuefeng committed
184
        repr_str = self.__class__.__name__
185
        repr_str += f'(sync_2d={self.sync_2d},'
186
        repr_str += f' flip_ratio_bev_vertical={self.flip_ratio_bev_vertical})'
wuyuefeng's avatar
wuyuefeng committed
187
        return repr_str
zhangwenwei's avatar
zhangwenwei committed
188

zhangwenwei's avatar
zhangwenwei committed
189

190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
@PIPELINES.register_module()
class RandomJitterPoints(object):
    """Randomly jitter point coordinates.

    Different from the global translation in ``GlobalRotScaleTrans``, here we \
        apply different noises to each point in a scene.

    Args:
        jitter_std (list[float]): The standard deviation of jittering noise.
            This applies random noise to all points in a 3D scene, which is \
            sampled from a gaussian distribution whose standard deviation is \
            set by ``jitter_std``. Defaults to [0.01, 0.01, 0.01]
        clip_range (list[float] | None): Clip the randomly generated jitter \
            noise into this range. If None is given, don't perform clipping.
            Defaults to [-0.05, 0.05]

    Note:
        This transform should only be used in point cloud segmentation tasks \
            because we don't transform ground-truth bboxes accordingly.
        For similar transform in detection task, please refer to `ObjectNoise`.
    """

    def __init__(self,
                 jitter_std=[0.01, 0.01, 0.01],
                 clip_range=[-0.05, 0.05]):
        seq_types = (list, tuple, np.ndarray)
        if not isinstance(jitter_std, seq_types):
            assert isinstance(jitter_std, (int, float)), \
                f'unsupported jitter_std type {type(jitter_std)}'
            jitter_std = [jitter_std, jitter_std, jitter_std]
        self.jitter_std = jitter_std

        if clip_range is not None:
            if not isinstance(clip_range, seq_types):
                assert isinstance(clip_range, (int, float)), \
                    f'unsupported clip_range type {type(clip_range)}'
                clip_range = [-clip_range, clip_range]
        self.clip_range = clip_range

    def __call__(self, input_dict):
        """Call function to jitter all the points in the scene.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
            dict: Results after adding noise to each point, \
                'points' key is updated in the result dict.
        """
        points = input_dict['points']
        jitter_std = np.array(self.jitter_std, dtype=np.float32)
        jitter_noise = \
            np.random.randn(points.shape[0], 3) * jitter_std[None, :]
        if self.clip_range is not None:
            jitter_noise = np.clip(jitter_noise, self.clip_range[0],
                                   self.clip_range[1])

        points.translate(jitter_noise)
        return input_dict

    def __repr__(self):
        """str: Return a string that describes the module."""
        repr_str = self.__class__.__name__
        repr_str += f'(jitter_std={self.jitter_std},'
        repr_str += f' clip_range={self.clip_range})'
        return repr_str


258
@PIPELINES.register_module()
zhangwenwei's avatar
zhangwenwei committed
259
class ObjectSample(object):
zhangwenwei's avatar
zhangwenwei committed
260
    """Sample GT objects to the data.
zhangwenwei's avatar
zhangwenwei committed
261
262
263
264
265

    Args:
        db_sampler (dict): Config dict of the database sampler.
        sample_2d (bool): Whether to also paste 2D image patch to the images
            This should be true when applying multi-modality cut-and-paste.
liyinhao's avatar
liyinhao committed
266
            Defaults to False.
zhangwenwei's avatar
zhangwenwei committed
267
    """
zhangwenwei's avatar
zhangwenwei committed
268
269
270
271
272
273
274
275
276
277

    def __init__(self, db_sampler, sample_2d=False):
        self.sampler_cfg = db_sampler
        self.sample_2d = sample_2d
        if 'type' not in db_sampler.keys():
            db_sampler['type'] = 'DataBaseSampler'
        self.db_sampler = build_from_cfg(db_sampler, OBJECTSAMPLERS)

    @staticmethod
    def remove_points_in_boxes(points, boxes):
278
279
280
        """Remove the points in the sampled bounding boxes.

        Args:
281
            points (:obj:`BasePoints`): Input point cloud array.
282
283
284
285
286
            boxes (np.ndarray): Sampled ground truth boxes.

        Returns:
            np.ndarray: Points with those in the boxes removed.
        """
287
        masks = box_np_ops.points_in_rbbox(points.coord.numpy(), boxes)
zhangwenwei's avatar
zhangwenwei committed
288
289
290
291
        points = points[np.logical_not(masks.any(-1))]
        return points

    def __call__(self, input_dict):
292
293
294
295
296
297
298
299
300
301
        """Call function to sample ground truth objects to the data.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
            dict: Results after object sampling augmentation, \
                'points', 'gt_bboxes_3d', 'gt_labels_3d' keys are updated \
                in the result dict.
        """
zhangwenwei's avatar
zhangwenwei committed
302
        gt_bboxes_3d = input_dict['gt_bboxes_3d']
zhangwenwei's avatar
zhangwenwei committed
303
304
        gt_labels_3d = input_dict['gt_labels_3d']

zhangwenwei's avatar
zhangwenwei committed
305
306
307
        # change to float for blending operation
        points = input_dict['points']
        if self.sample_2d:
wuyuefeng's avatar
wuyuefeng committed
308
            img = input_dict['img']
zhangwenwei's avatar
zhangwenwei committed
309
310
311
            gt_bboxes_2d = input_dict['gt_bboxes']
            # Assume for now 3D & 2D bboxes are the same
            sampled_dict = self.db_sampler.sample_all(
312
313
314
315
                gt_bboxes_3d.tensor.numpy(),
                gt_labels_3d,
                gt_bboxes_2d=gt_bboxes_2d,
                img=img)
zhangwenwei's avatar
zhangwenwei committed
316
317
        else:
            sampled_dict = self.db_sampler.sample_all(
318
                gt_bboxes_3d.tensor.numpy(), gt_labels_3d, img=None)
zhangwenwei's avatar
zhangwenwei committed
319
320
321
322

        if sampled_dict is not None:
            sampled_gt_bboxes_3d = sampled_dict['gt_bboxes_3d']
            sampled_points = sampled_dict['points']
zhangwenwei's avatar
zhangwenwei committed
323
            sampled_gt_labels = sampled_dict['gt_labels_3d']
zhangwenwei's avatar
zhangwenwei committed
324

zhangwenwei's avatar
zhangwenwei committed
325
326
            gt_labels_3d = np.concatenate([gt_labels_3d, sampled_gt_labels],
                                          axis=0)
327
328
329
            gt_bboxes_3d = gt_bboxes_3d.new_box(
                np.concatenate(
                    [gt_bboxes_3d.tensor.numpy(), sampled_gt_bboxes_3d]))
zhangwenwei's avatar
zhangwenwei committed
330

zhangwenwei's avatar
zhangwenwei committed
331
332
            points = self.remove_points_in_boxes(points, sampled_gt_bboxes_3d)
            # check the points dimension
333
            points = points.cat([sampled_points, points])
zhangwenwei's avatar
zhangwenwei committed
334
335
336
337
338

            if self.sample_2d:
                sampled_gt_bboxes_2d = sampled_dict['gt_bboxes_2d']
                gt_bboxes_2d = np.concatenate(
                    [gt_bboxes_2d, sampled_gt_bboxes_2d]).astype(np.float32)
zhangwenwei's avatar
zhangwenwei committed
339

zhangwenwei's avatar
zhangwenwei committed
340
                input_dict['gt_bboxes'] = gt_bboxes_2d
wuyuefeng's avatar
wuyuefeng committed
341
                input_dict['img'] = sampled_dict['img']
zhangwenwei's avatar
zhangwenwei committed
342
343

        input_dict['gt_bboxes_3d'] = gt_bboxes_3d
344
        input_dict['gt_labels_3d'] = gt_labels_3d.astype(np.long)
zhangwenwei's avatar
zhangwenwei committed
345
        input_dict['points'] = points
zhangwenwei's avatar
zhangwenwei committed
346

zhangwenwei's avatar
zhangwenwei committed
347
348
349
        return input_dict

    def __repr__(self):
350
        """str: Return a string that describes the module."""
351
352
353
354
355
356
357
358
359
        repr_str = self.__class__.__name__
        repr_str += f' sample_2d={self.sample_2d},'
        repr_str += f' data_root={self.sampler_cfg.data_root},'
        repr_str += f' info_path={self.sampler_cfg.info_path},'
        repr_str += f' rate={self.sampler_cfg.rate},'
        repr_str += f' prepare={self.sampler_cfg.prepare},'
        repr_str += f' classes={self.sampler_cfg.classes},'
        repr_str += f' sample_groups={self.sampler_cfg.sample_groups}'
        return repr_str
zhangwenwei's avatar
zhangwenwei committed
360
361


362
@PIPELINES.register_module()
zhangwenwei's avatar
zhangwenwei committed
363
class ObjectNoise(object):
zhangwenwei's avatar
zhangwenwei committed
364
    """Apply noise to each GT objects in the scene.
zhangwenwei's avatar
zhangwenwei committed
365
366

    Args:
367
        translation_std (list[float], optional): Standard deviation of the
zhangwenwei's avatar
zhangwenwei committed
368
369
            distribution where translation noise are sampled from.
            Defaults to [0.25, 0.25, 0.25].
370
        global_rot_range (list[float], optional): Global rotation to the scene.
zhangwenwei's avatar
zhangwenwei committed
371
            Defaults to [0.0, 0.0].
372
        rot_range (list[float], optional): Object rotation range.
zhangwenwei's avatar
zhangwenwei committed
373
374
375
376
            Defaults to [-0.15707963267, 0.15707963267].
        num_try (int, optional): Number of times to try if the noise applied is
            invalid. Defaults to 100.
    """
zhangwenwei's avatar
zhangwenwei committed
377
378

    def __init__(self,
zhangwenwei's avatar
zhangwenwei committed
379
                 translation_std=[0.25, 0.25, 0.25],
zhangwenwei's avatar
zhangwenwei committed
380
                 global_rot_range=[0.0, 0.0],
zhangwenwei's avatar
zhangwenwei committed
381
                 rot_range=[-0.15707963267, 0.15707963267],
zhangwenwei's avatar
zhangwenwei committed
382
                 num_try=100):
zhangwenwei's avatar
zhangwenwei committed
383
        self.translation_std = translation_std
zhangwenwei's avatar
zhangwenwei committed
384
        self.global_rot_range = global_rot_range
zhangwenwei's avatar
zhangwenwei committed
385
        self.rot_range = rot_range
zhangwenwei's avatar
zhangwenwei committed
386
387
388
        self.num_try = num_try

    def __call__(self, input_dict):
389
390
391
392
393
394
395
396
397
        """Call function to apply noise to each ground truth in the scene.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
            dict: Results after adding noise to each object, \
                'points', 'gt_bboxes_3d' keys are updated in the result dict.
        """
zhangwenwei's avatar
zhangwenwei committed
398
399
        gt_bboxes_3d = input_dict['gt_bboxes_3d']
        points = input_dict['points']
zhangwenwei's avatar
zhangwenwei committed
400

zhangwenwei's avatar
zhangwenwei committed
401
        # TODO: check this inplace function
402
        numpy_box = gt_bboxes_3d.tensor.numpy()
403
404
        numpy_points = points.tensor.numpy()

zhangwenwei's avatar
zhangwenwei committed
405
        noise_per_object_v3_(
406
            numpy_box,
407
            numpy_points,
zhangwenwei's avatar
zhangwenwei committed
408
409
            rotation_perturb=self.rot_range,
            center_noise_std=self.translation_std,
zhangwenwei's avatar
zhangwenwei committed
410
411
            global_random_rot_range=self.global_rot_range,
            num_try=self.num_try)
412
413

        input_dict['gt_bboxes_3d'] = gt_bboxes_3d.new_box(numpy_box)
414
        input_dict['points'] = points.new_point(numpy_points)
zhangwenwei's avatar
zhangwenwei committed
415
416
417
        return input_dict

    def __repr__(self):
418
        """str: Return a string that describes the module."""
zhangwenwei's avatar
zhangwenwei committed
419
        repr_str = self.__class__.__name__
420
421
422
423
        repr_str += f'(num_try={self.num_try},'
        repr_str += f' translation_std={self.translation_std},'
        repr_str += f' global_rot_range={self.global_rot_range},'
        repr_str += f' rot_range={self.rot_range})'
zhangwenwei's avatar
zhangwenwei committed
424
425
426
        return repr_str


427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
@PIPELINES.register_module()
class GlobalAlignment(object):
    """Apply global alignment to 3D scene points by rotation and translation.

    Args:
        rotation_axis (int): Rotation axis for points and bboxes rotation.

    Note:
        We do not record the applied rotation and translation as in \
            GlobalRotScaleTrans. Because usually, we do not need to reverse \
            the alignment step.
        For example, ScanNet 3D detection task uses aligned ground-truth \
            bounding boxes for evaluation.
    """

    def __init__(self, rotation_axis):
        self.rotation_axis = rotation_axis

    def _trans_points(self, input_dict, trans_factor):
        """Private function to translate points.

        Args:
            input_dict (dict): Result dict from loading pipeline.
            trans_factor (np.ndarray): Translation vector to be applied.

        Returns:
            dict: Results after translation, 'points' is updated in the dict.
        """
        input_dict['points'].translate(trans_factor)

    def _rot_points(self, input_dict, rot_mat):
        """Private function to rotate bounding boxes and points.

        Args:
            input_dict (dict): Result dict from loading pipeline.
            rot_mat (np.ndarray): Rotation matrix to be applied.

        Returns:
            dict: Results after rotation, 'points' is updated in the dict.
        """
        # input should be rot_mat_T so I transpose it here
        input_dict['points'].rotate(rot_mat.T)

    def _check_rot_mat(self, rot_mat):
        """Check if rotation matrix is valid for self.rotation_axis.

        Args:
            rot_mat (np.ndarray): Rotation matrix to be applied.
        """
        is_valid = np.allclose(np.linalg.det(rot_mat), 1.0)
        valid_array = np.zeros(3)
        valid_array[self.rotation_axis] = 1.0
        is_valid &= (rot_mat[self.rotation_axis, :] == valid_array).all()
        is_valid &= (rot_mat[:, self.rotation_axis] == valid_array).all()
        assert is_valid, f'invalid rotation matrix {rot_mat}'

    def __call__(self, input_dict):
        """Call function to shuffle points.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
            dict: Results after global alignment, 'points' and keys in \
                input_dict['bbox3d_fields'] are updated in the result dict.
        """
        assert 'axis_align_matrix' in input_dict['ann_info'].keys(), \
            'axis_align_matrix is not provided in GlobalAlignment'

        axis_align_matrix = input_dict['ann_info']['axis_align_matrix']
        assert axis_align_matrix.shape == (4, 4), \
            f'invalid shape {axis_align_matrix.shape} for axis_align_matrix'
        rot_mat = axis_align_matrix[:3, :3]
        trans_vec = axis_align_matrix[:3, -1]

        self._check_rot_mat(rot_mat)
        self._rot_points(input_dict, rot_mat)
        self._trans_points(input_dict, trans_vec)

        return input_dict

    def __repr__(self):
        repr_str = self.__class__.__name__
        repr_str += f'(rotation_axis={self.rotation_axis})'
        return repr_str


514
@PIPELINES.register_module()
zhangwenwei's avatar
zhangwenwei committed
515
class GlobalRotScaleTrans(object):
zhangwenwei's avatar
zhangwenwei committed
516
    """Apply global rotation, scaling and translation to a 3D scene.
zhangwenwei's avatar
zhangwenwei committed
517
518
519

    Args:
        rot_range (list[float]): Range of rotation angle.
liyinhao's avatar
liyinhao committed
520
            Defaults to [-0.78539816, 0.78539816] (close to [-pi/4, pi/4]).
zhangwenwei's avatar
zhangwenwei committed
521
        scale_ratio_range (list[float]): Range of scale ratio.
liyinhao's avatar
liyinhao committed
522
            Defaults to [0.95, 1.05].
523
524
        translation_std (list[float]): The standard deviation of translation
            noise. This applies random translation to a scene by a noise, which
zhangwenwei's avatar
zhangwenwei committed
525
            is sampled from a gaussian distribution whose standard deviation
liyinhao's avatar
liyinhao committed
526
527
            is set by ``translation_std``. Defaults to [0, 0, 0]
        shift_height (bool): Whether to shift height.
wuyuefeng's avatar
wuyuefeng committed
528
            (the fourth dimension of indoor points) when scaling.
liyinhao's avatar
liyinhao committed
529
            Defaults to False.
zhangwenwei's avatar
zhangwenwei committed
530
    """
zhangwenwei's avatar
zhangwenwei committed
531
532

    def __init__(self,
zhangwenwei's avatar
zhangwenwei committed
533
534
                 rot_range=[-0.78539816, 0.78539816],
                 scale_ratio_range=[0.95, 1.05],
wuyuefeng's avatar
wuyuefeng committed
535
536
                 translation_std=[0, 0, 0],
                 shift_height=False):
537
538
539
540
541
        seq_types = (list, tuple, np.ndarray)
        if not isinstance(rot_range, seq_types):
            assert isinstance(rot_range, (int, float)), \
                f'unsupported rot_range type {type(rot_range)}'
            rot_range = [-rot_range, rot_range]
zhangwenwei's avatar
zhangwenwei committed
542
        self.rot_range = rot_range
543
544
545

        assert isinstance(scale_ratio_range, seq_types), \
            f'unsupported scale_ratio_range type {type(scale_ratio_range)}'
zhangwenwei's avatar
zhangwenwei committed
546
        self.scale_ratio_range = scale_ratio_range
547
548
549
550
551
552
553

        if not isinstance(translation_std, seq_types):
            assert isinstance(translation_std, (int, float)), \
                f'unsupported translation_std type {type(translation_std)}'
            translation_std = [
                translation_std, translation_std, translation_std
            ]
554
555
        assert all([std >= 0 for std in translation_std]), \
            'translation_std should be positive'
zhangwenwei's avatar
zhangwenwei committed
556
        self.translation_std = translation_std
wuyuefeng's avatar
wuyuefeng committed
557
        self.shift_height = shift_height
zhangwenwei's avatar
zhangwenwei committed
558
559

    def _trans_bbox_points(self, input_dict):
560
561
562
563
564
565
566
567
568
569
        """Private function to translate bounding boxes and points.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
            dict: Results after translation, 'points', 'pcd_trans' \
                and keys in input_dict['bbox3d_fields'] are updated \
                in the result dict.
        """
570
        translation_std = np.array(self.translation_std, dtype=np.float32)
zhangwenwei's avatar
zhangwenwei committed
571
572
        trans_factor = np.random.normal(scale=translation_std, size=3).T

573
        input_dict['points'].translate(trans_factor)
zhangwenwei's avatar
zhangwenwei committed
574
575
576
577
578
        input_dict['pcd_trans'] = trans_factor
        for key in input_dict['bbox3d_fields']:
            input_dict[key].translate(trans_factor)

    def _rot_bbox_points(self, input_dict):
579
580
581
582
583
584
585
586
587
588
        """Private function to rotate bounding boxes and points.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
            dict: Results after rotation, 'points', 'pcd_rotation' \
                and keys in input_dict['bbox3d_fields'] are updated \
                in the result dict.
        """
zhangwenwei's avatar
zhangwenwei committed
589
        rotation = self.rot_range
zhangwenwei's avatar
zhangwenwei committed
590
        noise_rotation = np.random.uniform(rotation[0], rotation[1])
zhangwenwei's avatar
zhangwenwei committed
591

592
593
594
595
596
597
598
        # if no bbox in input_dict, only rotate points
        if len(input_dict['bbox3d_fields']) == 0:
            rot_mat_T = input_dict['points'].rotate(noise_rotation)
            input_dict['pcd_rotation'] = rot_mat_T
            return

        # rotate points with bboxes
zhangwenwei's avatar
zhangwenwei committed
599
        for key in input_dict['bbox3d_fields']:
wuyuefeng's avatar
wuyuefeng committed
600
601
602
603
604
            if len(input_dict[key].tensor) != 0:
                points, rot_mat_T = input_dict[key].rotate(
                    noise_rotation, input_dict['points'])
                input_dict['points'] = points
                input_dict['pcd_rotation'] = rot_mat_T
605

zhangwenwei's avatar
zhangwenwei committed
606
    def _scale_bbox_points(self, input_dict):
607
608
609
610
611
612
613
614
615
        """Private function to scale bounding boxes and points.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
            dict: Results after scaling, 'points'and keys in \
                input_dict['bbox3d_fields'] are updated in the result dict.
        """
zhangwenwei's avatar
zhangwenwei committed
616
        scale = input_dict['pcd_scale_factor']
617
618
        points = input_dict['points']
        points.scale(scale)
wuyuefeng's avatar
wuyuefeng committed
619
        if self.shift_height:
620
621
            assert 'height' in points.attribute_dims.keys(), \
                'setting shift_height=True but points have no height attribute'
622
623
            points.tensor[:, points.attribute_dims['height']] *= scale
        input_dict['points'] = points
wuyuefeng's avatar
wuyuefeng committed
624

zhangwenwei's avatar
zhangwenwei committed
625
626
        for key in input_dict['bbox3d_fields']:
            input_dict[key].scale(scale)
zhangwenwei's avatar
zhangwenwei committed
627

zhangwenwei's avatar
zhangwenwei committed
628
    def _random_scale(self, input_dict):
629
630
631
632
633
634
635
636
637
        """Private function to randomly set the scale factor.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
            dict: Results after scaling, 'pcd_scale_factor' are updated \
                in the result dict.
        """
zhangwenwei's avatar
zhangwenwei committed
638
639
640
        scale_factor = np.random.uniform(self.scale_ratio_range[0],
                                         self.scale_ratio_range[1])
        input_dict['pcd_scale_factor'] = scale_factor
zhangwenwei's avatar
zhangwenwei committed
641
642

    def __call__(self, input_dict):
643
644
645
646
647
648
649
650
651
652
653
        """Private function to rotate, scale and translate bounding boxes and \
        points.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
            dict: Results after scaling, 'points', 'pcd_rotation',
                'pcd_scale_factor', 'pcd_trans' and keys in \
                input_dict['bbox3d_fields'] are updated in the result dict.
        """
654
655
656
        if 'transformation_3d_flow' not in input_dict:
            input_dict['transformation_3d_flow'] = []

zhangwenwei's avatar
zhangwenwei committed
657
        self._rot_bbox_points(input_dict)
zhangwenwei's avatar
zhangwenwei committed
658

zhangwenwei's avatar
zhangwenwei committed
659
660
661
        if 'pcd_scale_factor' not in input_dict:
            self._random_scale(input_dict)
        self._scale_bbox_points(input_dict)
zhangwenwei's avatar
zhangwenwei committed
662

zhangwenwei's avatar
zhangwenwei committed
663
        self._trans_bbox_points(input_dict)
664
665

        input_dict['transformation_3d_flow'].extend(['R', 'S', 'T'])
zhangwenwei's avatar
zhangwenwei committed
666
667
668
        return input_dict

    def __repr__(self):
669
        """str: Return a string that describes the module."""
zhangwenwei's avatar
zhangwenwei committed
670
        repr_str = self.__class__.__name__
671
672
673
674
        repr_str += f'(rot_range={self.rot_range},'
        repr_str += f' scale_ratio_range={self.scale_ratio_range},'
        repr_str += f' translation_std={self.translation_std},'
        repr_str += f' shift_height={self.shift_height})'
zhangwenwei's avatar
zhangwenwei committed
675
676
677
        return repr_str


678
@PIPELINES.register_module()
zhangwenwei's avatar
zhangwenwei committed
679
class PointShuffle(object):
680
    """Shuffle input points."""
zhangwenwei's avatar
zhangwenwei committed
681
682

    def __call__(self, input_dict):
683
684
685
686
687
688
        """Call function to shuffle points.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
689
690
            dict: Results after filtering, 'points', 'pts_instance_mask' \
                and 'pts_semantic_mask' keys are updated in the result dict.
691
        """
692
693
694
695
696
697
698
699
700
701
702
703
        idx = input_dict['points'].shuffle()
        idx = idx.numpy()

        pts_instance_mask = input_dict.get('pts_instance_mask', None)
        pts_semantic_mask = input_dict.get('pts_semantic_mask', None)

        if pts_instance_mask is not None:
            input_dict['pts_instance_mask'] = pts_instance_mask[idx]

        if pts_semantic_mask is not None:
            input_dict['pts_semantic_mask'] = pts_semantic_mask[idx]

zhangwenwei's avatar
zhangwenwei committed
704
705
706
707
708
709
        return input_dict

    def __repr__(self):
        return self.__class__.__name__


710
@PIPELINES.register_module()
zhangwenwei's avatar
zhangwenwei committed
711
class ObjectRangeFilter(object):
712
713
714
715
716
    """Filter objects by the range.

    Args:
        point_cloud_range (list[float]): Point cloud range.
    """
zhangwenwei's avatar
zhangwenwei committed
717
718
719
720
721

    def __init__(self, point_cloud_range):
        self.pcd_range = np.array(point_cloud_range, dtype=np.float32)

    def __call__(self, input_dict):
722
723
724
725
726
727
728
729
730
        """Call function to filter objects by the range.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
            dict: Results after filtering, 'gt_bboxes_3d', 'gt_labels_3d' \
                keys are updated in the result dict.
        """
731
732
733
734
735
736
737
        # Check points instance type and initialise bev_range
        if isinstance(input_dict['gt_bboxes_3d'],
                      (LiDARInstance3DBoxes, DepthInstance3DBoxes)):
            bev_range = self.pcd_range[[0, 1, 3, 4]]
        elif isinstance(input_dict['gt_bboxes_3d'], CameraInstance3DBoxes):
            bev_range = self.pcd_range[[0, 2, 3, 5]]

zhangwenwei's avatar
zhangwenwei committed
738
        gt_bboxes_3d = input_dict['gt_bboxes_3d']
zhangwenwei's avatar
zhangwenwei committed
739
        gt_labels_3d = input_dict['gt_labels_3d']
740
        mask = gt_bboxes_3d.in_range_bev(bev_range)
zhangwenwei's avatar
zhangwenwei committed
741
        gt_bboxes_3d = gt_bboxes_3d[mask]
ZwwWayne's avatar
ZwwWayne committed
742
743
744
745
746
        # mask is a torch tensor but gt_labels_3d is still numpy array
        # using mask to index gt_labels_3d will cause bug when
        # len(gt_labels_3d) == 1, where mask=1 will be interpreted
        # as gt_labels_3d[1] and cause out of index error
        gt_labels_3d = gt_labels_3d[mask.numpy().astype(np.bool)]
zhangwenwei's avatar
zhangwenwei committed
747
748

        # limit rad to [-pi, pi]
749
750
        gt_bboxes_3d.limit_yaw(offset=0.5, period=2 * np.pi)
        input_dict['gt_bboxes_3d'] = gt_bboxes_3d
zhangwenwei's avatar
zhangwenwei committed
751
752
        input_dict['gt_labels_3d'] = gt_labels_3d

zhangwenwei's avatar
zhangwenwei committed
753
754
755
        return input_dict

    def __repr__(self):
756
        """str: Return a string that describes the module."""
zhangwenwei's avatar
zhangwenwei committed
757
        repr_str = self.__class__.__name__
758
        repr_str += f'(point_cloud_range={self.pcd_range.tolist()})'
zhangwenwei's avatar
zhangwenwei committed
759
760
761
        return repr_str


762
@PIPELINES.register_module()
zhangwenwei's avatar
zhangwenwei committed
763
class PointsRangeFilter(object):
764
765
766
767
768
    """Filter points by the range.

    Args:
        point_cloud_range (list[float]): Point cloud range.
    """
zhangwenwei's avatar
zhangwenwei committed
769
770

    def __init__(self, point_cloud_range):
771
        self.pcd_range = np.array(point_cloud_range, dtype=np.float32)
zhangwenwei's avatar
zhangwenwei committed
772
773

    def __call__(self, input_dict):
774
775
776
777
778
779
        """Call function to filter points by the range.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
780
781
            dict: Results after filtering, 'points', 'pts_instance_mask' \
                and 'pts_semantic_mask' keys are updated in the result dict.
782
        """
zhangwenwei's avatar
zhangwenwei committed
783
        points = input_dict['points']
784
785
        points_mask = points.in_range_3d(self.pcd_range)
        clean_points = points[points_mask]
zhangwenwei's avatar
zhangwenwei committed
786
        input_dict['points'] = clean_points
787
788
789
790
791
792
793
794
795
796
797
        points_mask = points_mask.numpy()

        pts_instance_mask = input_dict.get('pts_instance_mask', None)
        pts_semantic_mask = input_dict.get('pts_semantic_mask', None)

        if pts_instance_mask is not None:
            input_dict['pts_instance_mask'] = pts_instance_mask[points_mask]

        if pts_semantic_mask is not None:
            input_dict['pts_semantic_mask'] = pts_semantic_mask[points_mask]

zhangwenwei's avatar
zhangwenwei committed
798
799
800
        return input_dict

    def __repr__(self):
801
        """str: Return a string that describes the module."""
zhangwenwei's avatar
zhangwenwei committed
802
        repr_str = self.__class__.__name__
803
        repr_str += f'(point_cloud_range={self.pcd_range.tolist()})'
zhangwenwei's avatar
zhangwenwei committed
804
        return repr_str
zhangwenwei's avatar
zhangwenwei committed
805
806
807
808


@PIPELINES.register_module()
class ObjectNameFilter(object):
zhangwenwei's avatar
zhangwenwei committed
809
    """Filter GT objects by their names.
zhangwenwei's avatar
zhangwenwei committed
810
811

    Args:
liyinhao's avatar
liyinhao committed
812
        classes (list[str]): List of class names to be kept for training.
zhangwenwei's avatar
zhangwenwei committed
813
814
815
816
817
818
819
    """

    def __init__(self, classes):
        self.classes = classes
        self.labels = list(range(len(self.classes)))

    def __call__(self, input_dict):
820
821
822
823
824
825
826
827
828
        """Call function to filter objects by their names.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
            dict: Results after filtering, 'gt_bboxes_3d', 'gt_labels_3d' \
                keys are updated in the result dict.
        """
zhangwenwei's avatar
zhangwenwei committed
829
830
831
832
833
834
835
836
837
        gt_labels_3d = input_dict['gt_labels_3d']
        gt_bboxes_mask = np.array([n in self.labels for n in gt_labels_3d],
                                  dtype=np.bool_)
        input_dict['gt_bboxes_3d'] = input_dict['gt_bboxes_3d'][gt_bboxes_mask]
        input_dict['gt_labels_3d'] = input_dict['gt_labels_3d'][gt_bboxes_mask]

        return input_dict

    def __repr__(self):
838
        """str: Return a string that describes the module."""
zhangwenwei's avatar
zhangwenwei committed
839
840
841
        repr_str = self.__class__.__name__
        repr_str += f'(classes={self.classes})'
        return repr_str
wuyuefeng's avatar
wuyuefeng committed
842
843
844


@PIPELINES.register_module()
845
846
class PointSample(object):
    """Point sample.
wuyuefeng's avatar
wuyuefeng committed
847
848
849
850
851

    Sampling data to a certain number.

    Args:
        num_points (int): Number of points to be sampled.
852
        sample_range (float, optional): The range where to sample points.
853
854
855
856
            If not None, the points with depth larger than `sample_range` are
            prior to be sampled. Defaults to None.
        replace (bool, optional): Whether the sampling is with or without
            replacement. Defaults to False.
wuyuefeng's avatar
wuyuefeng committed
857
858
    """

859
    def __init__(self, num_points, sample_range=None, replace=False):
wuyuefeng's avatar
wuyuefeng committed
860
        self.num_points = num_points
861
862
863
864
865
866
867
868
869
        self.sample_range = sample_range
        self.replace = replace

    def _points_random_sampling(self,
                                points,
                                num_samples,
                                sample_range=None,
                                replace=False,
                                return_choices=False):
wuyuefeng's avatar
wuyuefeng committed
870
871
872
873
874
        """Points random sampling.

        Sample points to a certain number.

        Args:
875
            points (np.ndarray | :obj:`BasePoints`): 3D Points.
wuyuefeng's avatar
wuyuefeng committed
876
            num_samples (int): Number of samples to be sampled.
877
            sample_range (float, optional): Indicating the range where the
878
                points will be sampled. Defaults to None.
879
880
881
882
            replace (bool, optional): Sampling with or without replacement.
                Defaults to None.
            return_choices (bool, optional): Whether return choice.
                Defaults to False.
wuyuefeng's avatar
wuyuefeng committed
883
        Returns:
884
            tuple[np.ndarray] | np.ndarray:
885
                - points (np.ndarray | :obj:`BasePoints`): 3D Points.
886
                - choices (np.ndarray, optional): The generated random samples.
wuyuefeng's avatar
wuyuefeng committed
887
        """
888
        if not replace:
wuyuefeng's avatar
wuyuefeng committed
889
            replace = (points.shape[0] < num_samples)
890
891
892
893
894
895
        point_range = range(len(points))
        if sample_range is not None and not replace:
            # Only sampling the near points when len(points) >= num_samples
            depth = np.linalg.norm(points.tensor, axis=1)
            far_inds = np.where(depth > sample_range)[0]
            near_inds = np.where(depth <= sample_range)[0]
896
897
898
899
            # in case there are too many far points
            if len(far_inds) > num_samples:
                far_inds = np.random.choice(
                    far_inds, num_samples, replace=False)
900
901
902
903
904
905
906
            point_range = near_inds
            num_samples -= len(far_inds)
        choices = np.random.choice(point_range, num_samples, replace=replace)
        if sample_range is not None and not replace:
            choices = np.concatenate((far_inds, choices))
            # Shuffle points after sampling
            np.random.shuffle(choices)
wuyuefeng's avatar
wuyuefeng committed
907
908
909
910
911
912
        if return_choices:
            return points[choices], choices
        else:
            return points[choices]

    def __call__(self, results):
913
914
915
916
917
918
919
920
        """Call function to sample points to in indoor scenes.

        Args:
            input_dict (dict): Result dict from loading pipeline.
        Returns:
            dict: Results after sampling, 'points', 'pts_instance_mask' \
                and 'pts_semantic_mask' keys are updated in the result dict.
        """
wuyuefeng's avatar
wuyuefeng committed
921
        points = results['points']
922
        # Points in Camera coord can provide the depth information.
923
        # TODO: Need to support distance-based sampling for other coord system.
924
        if self.sample_range is not None:
925
            from mmdet3d.core.points import CameraPoints
926
            assert isinstance(points, CameraPoints), \
927
                'Sampling based on distance is only applicable for CAM coord'
928
929
930
931
932
933
        points, choices = self._points_random_sampling(
            points,
            self.num_points,
            self.sample_range,
            self.replace,
            return_choices=True)
934
        results['points'] = points
935

wuyuefeng's avatar
wuyuefeng committed
936
937
938
        pts_instance_mask = results.get('pts_instance_mask', None)
        pts_semantic_mask = results.get('pts_semantic_mask', None)

939
        if pts_instance_mask is not None:
wuyuefeng's avatar
wuyuefeng committed
940
941
            pts_instance_mask = pts_instance_mask[choices]
            results['pts_instance_mask'] = pts_instance_mask
942
943
944

        if pts_semantic_mask is not None:
            pts_semantic_mask = pts_semantic_mask[choices]
wuyuefeng's avatar
wuyuefeng committed
945
946
947
948
949
            results['pts_semantic_mask'] = pts_semantic_mask

        return results

    def __repr__(self):
950
        """str: Return a string that describes the module."""
wuyuefeng's avatar
wuyuefeng committed
951
        repr_str = self.__class__.__name__
952
        repr_str += f'(num_points={self.num_points},'
953
954
        repr_str += f' sample_range={self.sample_range},'
        repr_str += f' replace={self.replace})'
955

956
957
958
        return repr_str


959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
@PIPELINES.register_module()
class IndoorPointSample(PointSample):
    """Indoor point sample.

    Sampling data to a certain number.
    NOTE: IndoorPointSample is deprecated in favor of PointSample

    Args:
        num_points (int): Number of points to be sampled.
    """

    def __init__(self, *args, **kwargs):
        warnings.warn(
            'IndoorPointSample is deprecated in favor of PointSample')
        super(IndoorPointSample, self).__init__(*args, **kwargs)


976
977
978
979
980
981
982
983
984
985
986
987
@PIPELINES.register_module()
class IndoorPatchPointSample(object):
    r"""Indoor point sample within a patch. Modified from `PointNet++ <https://
    github.com/charlesq34/pointnet2/blob/master/scannet/scannet_dataset.py>`_.

    Sampling data to a certain number for semantic segmentation.

    Args:
        num_points (int): Number of points to be sampled.
        block_size (float, optional): Size of a block to sample points from.
            Defaults to 1.5.
        sample_rate (float, optional): Stride used in sliding patch generation.
988
989
990
            This parameter is unused in `IndoorPatchPointSample` and thus has
            been deprecated. We plan to remove it in the future.
            Defaults to None.
991
992
        ignore_index (int, optional): Label index that won't be used for the
            segmentation task. This is set in PointSegClassMapping as neg_cls.
993
            If not None, will be used as a patch selection criterion.
994
995
996
997
998
            Defaults to None.
        use_normalized_coord (bool, optional): Whether to use normalized xyz as
            additional features. Defaults to False.
        num_try (int, optional): Number of times to try if the patch selected
            is invalid. Defaults to 10.
999
1000
        enlarge_size (float | None, optional): Enlarge the sampled patch to
            [-block_size / 2 - enlarge_size, block_size / 2 + enlarge_size] as
1001
            an augmentation. If None, set it as 0. Defaults to 0.2.
1002
1003
1004
        min_unique_num (int | None, optional): Minimum number of unique points
            the sampled patch should contain. If None, use PointNet++'s method
            to judge uniqueness. Defaults to None.
1005
1006
        eps (float, optional): A value added to patch boundary to guarantee
            points coverage. Defaults to 1e-2.
1007
1008
1009
1010
1011
1012

    Note:
        This transform should only be used in the training process of point
            cloud segmentation tasks. For the sliding patch generation and
            inference process in testing, please refer to the `slide_inference`
            function of `EncoderDecoder3D` class.
1013
1014
1015
1016
1017
    """

    def __init__(self,
                 num_points,
                 block_size=1.5,
1018
                 sample_rate=None,
1019
1020
                 ignore_index=None,
                 use_normalized_coord=False,
1021
1022
                 num_try=10,
                 enlarge_size=0.2,
1023
1024
                 min_unique_num=None,
                 eps=1e-2):
1025
1026
1027
1028
1029
        self.num_points = num_points
        self.block_size = block_size
        self.ignore_index = ignore_index
        self.use_normalized_coord = use_normalized_coord
        self.num_try = num_try
1030
        self.enlarge_size = enlarge_size if enlarge_size is not None else 0.0
1031
        self.min_unique_num = min_unique_num
1032
        self.eps = eps
1033
1034
1035
1036
1037

        if sample_rate is not None:
            warnings.warn(
                "'sample_rate' has been deprecated and will be removed in "
                'the future. Please remove them from your code.')
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052

    def _input_generation(self, coords, patch_center, coord_max, attributes,
                          attribute_dims, point_type):
        """Generating model input.

        Generate input by subtracting patch center and adding additional \
            features. Currently support colors and normalized xyz as features.

        Args:
            coords (np.ndarray): Sampled 3D Points.
            patch_center (np.ndarray): Center coordinate of the selected patch.
            coord_max (np.ndarray): Max coordinate of all 3D Points.
            attributes (np.ndarray): features of input points.
            attribute_dims (dict): Dictionary to indicate the meaning of extra
                dimension.
1053
            point_type (type): class of input points inherited from BasePoints.
1054
1055

        Returns:
1056
            :obj:`BasePoints`: The generated input data.
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
        """
        # subtract patch center, the z dimension is not centered
        centered_coords = coords.copy()
        centered_coords[:, 0] -= patch_center[0]
        centered_coords[:, 1] -= patch_center[1]

        if self.use_normalized_coord:
            normalized_coord = coords / coord_max
            attributes = np.concatenate([attributes, normalized_coord], axis=1)
            if attribute_dims is None:
                attribute_dims = dict()
            attribute_dims.update(
                dict(normalized_coord=[
                    attributes.shape[1], attributes.shape[1] +
                    1, attributes.shape[1] + 2
                ]))

        points = np.concatenate([centered_coords, attributes], axis=1)
        points = point_type(
            points, points_dim=points.shape[1], attribute_dims=attribute_dims)

        return points

1080
    def _patch_points_sampling(self, points, sem_mask):
1081
1082
1083
1084
1085
1086
        """Patch points sampling.

        First sample a valid patch.
        Then sample points within that patch to a certain number.

        Args:
1087
            points (:obj:`BasePoints`): 3D Points.
1088
1089
1090
            sem_mask (np.ndarray): semantic segmentation mask for input points.

        Returns:
1091
            tuple[:obj:`BasePoints`, np.ndarray] | :obj:`BasePoints`:
1092

1093
                - points (:obj:`BasePoints`): 3D Points.
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
                - choices (np.ndarray): The generated random samples.
        """
        coords = points.coord.numpy()
        attributes = points.tensor[:, 3:].numpy()
        attribute_dims = points.attribute_dims
        point_type = type(points)

        coord_max = np.amax(coords, axis=0)
        coord_min = np.amin(coords, axis=0)

1104
        for _ in range(self.num_try):
1105
1106
1107
            # random sample a point as patch center
            cur_center = coords[np.random.choice(coords.shape[0])]

1108
1109
            # boundary of a patch, which would be enlarged by
            # `self.enlarge_size` as an augmentation
1110
1111
1112
1113
1114
1115
1116
            cur_max = cur_center + np.array(
                [self.block_size / 2.0, self.block_size / 2.0, 0.0])
            cur_min = cur_center - np.array(
                [self.block_size / 2.0, self.block_size / 2.0, 0.0])
            cur_max[2] = coord_max[2]
            cur_min[2] = coord_min[2]
            cur_choice = np.sum(
1117
1118
                (coords >= (cur_min - self.enlarge_size)) *
                (coords <= (cur_max + self.enlarge_size)),
1119
1120
1121
1122
1123
1124
1125
                axis=1) == 3

            if not cur_choice.any():  # no points in this patch
                continue

            cur_coords = coords[cur_choice, :]
            cur_sem_mask = sem_mask[cur_choice]
1126
            point_idxs = np.where(cur_choice)[0]
1127
            mask = np.sum(
1128
1129
                (cur_coords >= (cur_min - self.eps)) * (cur_coords <=
                                                        (cur_max + self.eps)),
1130
                axis=1) == 3
1131

1132
1133
            # two criteria for patch sampling, adopted from PointNet++
            # 1. selected patch should contain enough unique points
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
            if self.min_unique_num is None:
                # use PointNet++'s method as default
                # [31, 31, 62] are just some big values used to transform
                # coords from 3d array to 1d and then check their uniqueness
                # this is used in all the ScanNet code following PointNet++
                vidx = np.ceil(
                    (cur_coords[mask, :] - cur_min) / (cur_max - cur_min) *
                    np.array([31.0, 31.0, 62.0]))
                vidx = np.unique(vidx[:, 0] * 31.0 * 62.0 + vidx[:, 1] * 62.0 +
                                 vidx[:, 2])
                flag1 = len(vidx) / 31.0 / 31.0 / 62.0 >= 0.02
            else:
1146
                # if `min_unique_num` is provided, directly compare with it
1147
                flag1 = mask.sum() >= self.min_unique_num
1148

1149
            # 2. selected patch should contain enough annotated points
1150
1151
1152
1153
1154
1155
1156
1157
1158
            if self.ignore_index is None:
                flag2 = True
            else:
                flag2 = np.sum(cur_sem_mask != self.ignore_index) / \
                               len(cur_sem_mask) >= 0.7

            if flag1 and flag2:
                break

1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
        # sample idx to `self.num_points`
        if point_idxs.size >= self.num_points:
            # no duplicate in sub-sampling
            choices = np.random.choice(
                point_idxs, self.num_points, replace=False)
        else:
            # do not use random choice here to avoid some points not counted
            dup = np.random.choice(point_idxs.size,
                                   self.num_points - point_idxs.size)
            idx_dup = np.concatenate(
                [np.arange(point_idxs.size),
                 np.array(dup)], 0)
            choices = point_idxs[idx_dup]
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213

        # construct model input
        points = self._input_generation(coords[choices], cur_center, coord_max,
                                        attributes[choices], attribute_dims,
                                        point_type)

        return points, choices

    def __call__(self, results):
        """Call function to sample points to in indoor scenes.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
            dict: Results after sampling, 'points', 'pts_instance_mask' \
                and 'pts_semantic_mask' keys are updated in the result dict.
        """
        points = results['points']

        assert 'pts_semantic_mask' in results.keys(), \
            'semantic mask should be provided in training and evaluation'
        pts_semantic_mask = results['pts_semantic_mask']

        points, choices = self._patch_points_sampling(points,
                                                      pts_semantic_mask)

        results['points'] = points
        results['pts_semantic_mask'] = pts_semantic_mask[choices]
        pts_instance_mask = results.get('pts_instance_mask', None)
        if pts_instance_mask is not None:
            results['pts_instance_mask'] = pts_instance_mask[choices]

        return results

    def __repr__(self):
        """str: Return a string that describes the module."""
        repr_str = self.__class__.__name__
        repr_str += f'(num_points={self.num_points},'
        repr_str += f' block_size={self.block_size},'
        repr_str += f' ignore_index={self.ignore_index},'
        repr_str += f' use_normalized_coord={self.use_normalized_coord},'
1214
1215
        repr_str += f' num_try={self.num_try},'
        repr_str += f' enlarge_size={self.enlarge_size},'
1216
1217
        repr_str += f' min_unique_num={self.min_unique_num},'
        repr_str += f' eps={self.eps})'
wuyuefeng's avatar
wuyuefeng committed
1218
        return repr_str
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246


@PIPELINES.register_module()
class BackgroundPointsFilter(object):
    """Filter background points near the bounding box.

    Args:
        bbox_enlarge_range (tuple[float], float): Bbox enlarge range.
    """

    def __init__(self, bbox_enlarge_range):
        assert (is_tuple_of(bbox_enlarge_range, float)
                and len(bbox_enlarge_range) == 3) \
            or isinstance(bbox_enlarge_range, float), \
            f'Invalid arguments bbox_enlarge_range {bbox_enlarge_range}'

        if isinstance(bbox_enlarge_range, float):
            bbox_enlarge_range = [bbox_enlarge_range] * 3
        self.bbox_enlarge_range = np.array(
            bbox_enlarge_range, dtype=np.float32)[np.newaxis, :]

    def __call__(self, input_dict):
        """Call function to filter points by the range.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
1247
1248
            dict: Results after filtering, 'points', 'pts_instance_mask' \
                and 'pts_semantic_mask' keys are updated in the result dict.
1249
1250
1251
1252
        """
        points = input_dict['points']
        gt_bboxes_3d = input_dict['gt_bboxes_3d']

xiliu8006's avatar
xiliu8006 committed
1253
1254
1255
1256
        # avoid groundtruth being modified
        gt_bboxes_3d_np = gt_bboxes_3d.tensor.clone().numpy()
        gt_bboxes_3d_np[:, :3] = gt_bboxes_3d.gravity_center.clone().numpy()

1257
1258
        enlarged_gt_bboxes_3d = gt_bboxes_3d_np.copy()
        enlarged_gt_bboxes_3d[:, 3:6] += self.bbox_enlarge_range
xiliu8006's avatar
xiliu8006 committed
1259
        points_numpy = points.tensor.clone().numpy()
1260
1261
        foreground_masks = box_np_ops.points_in_rbbox(
            points_numpy, gt_bboxes_3d_np, origin=(0.5, 0.5, 0.5))
1262
        enlarge_foreground_masks = box_np_ops.points_in_rbbox(
1263
            points_numpy, enlarged_gt_bboxes_3d, origin=(0.5, 0.5, 0.5))
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
        foreground_masks = foreground_masks.max(1)
        enlarge_foreground_masks = enlarge_foreground_masks.max(1)
        valid_masks = ~np.logical_and(~foreground_masks,
                                      enlarge_foreground_masks)

        input_dict['points'] = points[valid_masks]
        pts_instance_mask = input_dict.get('pts_instance_mask', None)
        if pts_instance_mask is not None:
            input_dict['pts_instance_mask'] = pts_instance_mask[valid_masks]

        pts_semantic_mask = input_dict.get('pts_semantic_mask', None)
        if pts_semantic_mask is not None:
            input_dict['pts_semantic_mask'] = pts_semantic_mask[valid_masks]
        return input_dict

    def __repr__(self):
        """str: Return a string that describes the module."""
        repr_str = self.__class__.__name__
1282
        repr_str += f'(bbox_enlarge_range={self.bbox_enlarge_range.tolist()})'
1283
        return repr_str
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294


@PIPELINES.register_module()
class VoxelBasedPointSampler(object):
    """Voxel based point sampler.

    Apply voxel sampling to multiple sweep points.

    Args:
        cur_sweep_cfg (dict): Config for sampling current points.
        prev_sweep_cfg (dict): Config for sampling previous points.
1295
        time_dim (int): Index that indicate the time dimension
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
            for input points.
    """

    def __init__(self, cur_sweep_cfg, prev_sweep_cfg=None, time_dim=3):
        self.cur_voxel_generator = VoxelGenerator(**cur_sweep_cfg)
        self.cur_voxel_num = self.cur_voxel_generator._max_voxels
        self.time_dim = time_dim
        if prev_sweep_cfg is not None:
            assert prev_sweep_cfg['max_num_points'] == \
                cur_sweep_cfg['max_num_points']
            self.prev_voxel_generator = VoxelGenerator(**prev_sweep_cfg)
            self.prev_voxel_num = self.prev_voxel_generator._max_voxels
        else:
            self.prev_voxel_generator = None
            self.prev_voxel_num = 0

    def _sample_points(self, points, sampler, point_dim):
        """Sample points for each points subset.

        Args:
            points (np.ndarray): Points subset to be sampled.
            sampler (VoxelGenerator): Voxel based sampler for
                each points subset.
1319
            point_dim (int): The dimension of each points
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355

        Returns:
            np.ndarray: Sampled points.
        """
        voxels, coors, num_points_per_voxel = sampler.generate(points)
        if voxels.shape[0] < sampler._max_voxels:
            padding_points = np.zeros([
                sampler._max_voxels - voxels.shape[0], sampler._max_num_points,
                point_dim
            ],
                                      dtype=points.dtype)
            padding_points[:] = voxels[0]
            sample_points = np.concatenate([voxels, padding_points], axis=0)
        else:
            sample_points = voxels

        return sample_points

    def __call__(self, results):
        """Call function to sample points from multiple sweeps.

        Args:
            input_dict (dict): Result dict from loading pipeline.

        Returns:
            dict: Results after sampling, 'points', 'pts_instance_mask' \
                and 'pts_semantic_mask' keys are updated in the result dict.
        """
        points = results['points']
        original_dim = points.shape[1]

        # TODO: process instance and semantic mask while _max_num_points
        # is larger than 1
        # Extend points with seg and mask fields
        map_fields2dim = []
        start_dim = original_dim
1356
1357
        points_numpy = points.tensor.numpy()
        extra_channel = [points_numpy]
1358
1359
1360
1361
1362
1363
1364
1365
1366
        for idx, key in enumerate(results['pts_mask_fields']):
            map_fields2dim.append((key, idx + start_dim))
            extra_channel.append(results[key][..., None])

        start_dim += len(results['pts_mask_fields'])
        for idx, key in enumerate(results['pts_seg_fields']):
            map_fields2dim.append((key, idx + start_dim))
            extra_channel.append(results[key][..., None])

1367
        points_numpy = np.concatenate(extra_channel, axis=-1)
1368
1369
1370
1371
1372

        # Split points into two part, current sweep points and
        # previous sweeps points.
        # TODO: support different sampling methods for next sweeps points
        # and previous sweeps points.
1373
1374
1375
        cur_points_flag = (points_numpy[:, self.time_dim] == 0)
        cur_sweep_points = points_numpy[cur_points_flag]
        prev_sweeps_points = points_numpy[~cur_points_flag]
1376
1377
1378
1379
1380
1381
1382
1383
1384
        if prev_sweeps_points.shape[0] == 0:
            prev_sweeps_points = cur_sweep_points

        # Shuffle points before sampling
        np.random.shuffle(cur_sweep_points)
        np.random.shuffle(prev_sweeps_points)

        cur_sweep_points = self._sample_points(cur_sweep_points,
                                               self.cur_voxel_generator,
1385
                                               points_numpy.shape[1])
1386
1387
1388
        if self.prev_voxel_generator is not None:
            prev_sweeps_points = self._sample_points(prev_sweeps_points,
                                                     self.prev_voxel_generator,
1389
                                                     points_numpy.shape[1])
1390

1391
1392
            points_numpy = np.concatenate(
                [cur_sweep_points, prev_sweeps_points], 0)
1393
        else:
1394
            points_numpy = cur_sweep_points
1395
1396

        if self.cur_voxel_generator._max_num_points == 1:
1397
1398
            points_numpy = points_numpy.squeeze(1)
        results['points'] = points.new_point(points_numpy[..., :original_dim])
1399

1400
        # Restore the corresponding seg and mask fields
1401
        for key, dim_index in map_fields2dim:
1402
            results[key] = points_numpy[..., dim_index]
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425

        return results

    def __repr__(self):
        """str: Return a string that describes the module."""

        def _auto_indent(repr_str, indent):
            repr_str = repr_str.split('\n')
            repr_str = [' ' * indent + t + '\n' for t in repr_str]
            repr_str = ''.join(repr_str)[:-1]
            return repr_str

        repr_str = self.__class__.__name__
        indent = 4
        repr_str += '(\n'
        repr_str += ' ' * indent + f'num_cur_sweep={self.cur_voxel_num},\n'
        repr_str += ' ' * indent + f'num_prev_sweep={self.prev_voxel_num},\n'
        repr_str += ' ' * indent + f'time_dim={self.time_dim},\n'
        repr_str += ' ' * indent + 'cur_voxel_generator=\n'
        repr_str += f'{_auto_indent(repr(self.cur_voxel_generator), 8)},\n'
        repr_str += ' ' * indent + 'prev_voxel_generator=\n'
        repr_str += f'{_auto_indent(repr(self.prev_voxel_generator), 8)})'
        return repr_str