transforms_3d.py 22.8 KB
Newer Older
zhangwenwei's avatar
zhangwenwei committed
1
import numpy as np
2
from mmcv.utils import build_from_cfg
zhangwenwei's avatar
zhangwenwei committed
3
4

from mmdet3d.core.bbox import box_np_ops
5
from mmdet.datasets.builder import PIPELINES
zhangwenwei's avatar
zhangwenwei committed
6
from mmdet.datasets.pipelines import RandomFlip
zhangwenwei's avatar
zhangwenwei committed
7
8
9
10
from ..registry import OBJECTSAMPLERS
from .data_augment_utils import noise_per_object_v3_


11
@PIPELINES.register_module()
zhangwenwei's avatar
zhangwenwei committed
12
13
14
15
16
17
18
19
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
20
21
22
        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
23
            to that of 2D images. Defaults to True.
wuyuefeng's avatar
wuyuefeng committed
24
        flip_ratio_bev_horizontal (float, optional): The flipping probability
liyinhao's avatar
liyinhao committed
25
            in horizontal direction. Defaults to 0.0.
wuyuefeng's avatar
wuyuefeng committed
26
        flip_ratio_bev_vertical (float, optional): The flipping probability
liyinhao's avatar
liyinhao committed
27
            in vertical direction. Defaults to 0.0.
zhangwenwei's avatar
zhangwenwei committed
28
29
    """

wuyuefeng's avatar
wuyuefeng committed
30
31
32
33
34
35
36
    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
37
        self.sync_2d = sync_2d
wuyuefeng's avatar
wuyuefeng committed
38
39
40
41
42
43
44
45
46
47
48
        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'):
49
50
51
52
53
54
55
56
57
58
        """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
59
        assert direction in ['horizontal', 'vertical']
60
61
62
63
64
        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
65
        for key in input_dict['bbox3d_fields']:
wuyuefeng's avatar
wuyuefeng committed
66
67
            input_dict['points'] = input_dict[key].flip(
                direction, points=input_dict['points'])
zhangwenwei's avatar
zhangwenwei committed
68
69

    def __call__(self, input_dict):
70
71
72
73
74
75
76
77
78
79
80
        """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.
        """
zhangwenwei's avatar
zhangwenwei committed
81
        # filp 2D image and its annotations
zhangwenwei's avatar
zhangwenwei committed
82
        super(RandomFlip3D, self).__call__(input_dict)
zhangwenwei's avatar
zhangwenwei committed
83

zhangwenwei's avatar
zhangwenwei committed
84
        if self.sync_2d:
wuyuefeng's avatar
wuyuefeng committed
85
86
            input_dict['pcd_horizontal_flip'] = input_dict['flip']
            input_dict['pcd_vertical_flip'] = False
zhangwenwei's avatar
zhangwenwei committed
87
        else:
wuyuefeng's avatar
wuyuefeng committed
88
89
90
91
92
93
94
95
96
97
98
99
100
            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

        if input_dict['pcd_horizontal_flip']:
            self.random_flip_data_3d(input_dict, 'horizontal')
        if input_dict['pcd_vertical_flip']:
            self.random_flip_data_3d(input_dict, 'vertical')
zhangwenwei's avatar
zhangwenwei committed
101
102
        return input_dict

zhangwenwei's avatar
zhangwenwei committed
103
    def __repr__(self):
104
        """str: Return a string that describes the module."""
wuyuefeng's avatar
wuyuefeng committed
105
106
107
108
109
110
111
        repr_str = self.__class__.__name__
        repr_str += '(sync_2d={},'.format(self.sync_2d)
        repr_str += '(flip_ratio_bev_horizontal={},'.format(
            self.flip_ratio_bev_horizontal)
        repr_str += '(flip_ratio_bev_vertical={},'.format(
            self.flip_ratio_bev_vertical)
        return repr_str
zhangwenwei's avatar
zhangwenwei committed
112

zhangwenwei's avatar
zhangwenwei committed
113

114
@PIPELINES.register_module()
zhangwenwei's avatar
zhangwenwei committed
115
class ObjectSample(object):
zhangwenwei's avatar
zhangwenwei committed
116
    """Sample GT objects to the data.
zhangwenwei's avatar
zhangwenwei committed
117
118
119
120
121

    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
122
            Defaults to False.
zhangwenwei's avatar
zhangwenwei committed
123
    """
zhangwenwei's avatar
zhangwenwei committed
124
125
126
127
128
129
130
131
132
133

    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):
134
135
136
137
138
139
140
141
142
        """Remove the points in the sampled bounding boxes.

        Args:
            points (np.ndarray): Input point cloud array.
            boxes (np.ndarray): Sampled ground truth boxes.

        Returns:
            np.ndarray: Points with those in the boxes removed.
        """
zhangwenwei's avatar
zhangwenwei committed
143
144
145
146
147
        masks = box_np_ops.points_in_rbbox(points, boxes)
        points = points[np.logical_not(masks.any(-1))]
        return points

    def __call__(self, input_dict):
148
149
150
151
152
153
154
155
156
157
        """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
158
        gt_bboxes_3d = input_dict['gt_bboxes_3d']
zhangwenwei's avatar
zhangwenwei committed
159
160
        gt_labels_3d = input_dict['gt_labels_3d']

zhangwenwei's avatar
zhangwenwei committed
161
162
163
        # change to float for blending operation
        points = input_dict['points']
        if self.sample_2d:
wuyuefeng's avatar
wuyuefeng committed
164
            img = input_dict['img']
zhangwenwei's avatar
zhangwenwei committed
165
166
167
            gt_bboxes_2d = input_dict['gt_bboxes']
            # Assume for now 3D & 2D bboxes are the same
            sampled_dict = self.db_sampler.sample_all(
168
169
170
171
                gt_bboxes_3d.tensor.numpy(),
                gt_labels_3d,
                gt_bboxes_2d=gt_bboxes_2d,
                img=img)
zhangwenwei's avatar
zhangwenwei committed
172
173
        else:
            sampled_dict = self.db_sampler.sample_all(
174
                gt_bboxes_3d.tensor.numpy(), gt_labels_3d, img=None)
zhangwenwei's avatar
zhangwenwei committed
175
176
177
178

        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
179
            sampled_gt_labels = sampled_dict['gt_labels_3d']
zhangwenwei's avatar
zhangwenwei committed
180

zhangwenwei's avatar
zhangwenwei committed
181
182
            gt_labels_3d = np.concatenate([gt_labels_3d, sampled_gt_labels],
                                          axis=0)
183
184
185
            gt_bboxes_3d = gt_bboxes_3d.new_box(
                np.concatenate(
                    [gt_bboxes_3d.tensor.numpy(), sampled_gt_bboxes_3d]))
zhangwenwei's avatar
zhangwenwei committed
186

zhangwenwei's avatar
zhangwenwei committed
187
188
189
190
191
192
193
194
195
196
            points = self.remove_points_in_boxes(points, sampled_gt_bboxes_3d)
            # check the points dimension
            dim_inds = points.shape[-1]
            points = np.concatenate([sampled_points[:, :dim_inds], points],
                                    axis=0)

            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
197

zhangwenwei's avatar
zhangwenwei committed
198
                input_dict['gt_bboxes'] = gt_bboxes_2d
wuyuefeng's avatar
wuyuefeng committed
199
                input_dict['img'] = sampled_dict['img']
zhangwenwei's avatar
zhangwenwei committed
200
201

        input_dict['gt_bboxes_3d'] = gt_bboxes_3d
zhangwenwei's avatar
zhangwenwei committed
202
        input_dict['gt_labels_3d'] = gt_labels_3d
zhangwenwei's avatar
zhangwenwei committed
203
        input_dict['points'] = points
zhangwenwei's avatar
zhangwenwei committed
204

zhangwenwei's avatar
zhangwenwei committed
205
206
207
        return input_dict

    def __repr__(self):
208
        """str: Return a string that describes the module."""
zhangwenwei's avatar
zhangwenwei committed
209
210
211
        return self.__class__.__name__


212
@PIPELINES.register_module()
zhangwenwei's avatar
zhangwenwei committed
213
class ObjectNoise(object):
zhangwenwei's avatar
zhangwenwei committed
214
    """Apply noise to each GT objects in the scene.
zhangwenwei's avatar
zhangwenwei committed
215
216

    Args:
217
        translation_std (list[float], optional): Standard deviation of the
zhangwenwei's avatar
zhangwenwei committed
218
219
            distribution where translation noise are sampled from.
            Defaults to [0.25, 0.25, 0.25].
220
        global_rot_range (list[float], optional): Global rotation to the scene.
zhangwenwei's avatar
zhangwenwei committed
221
            Defaults to [0.0, 0.0].
222
        rot_range (list[float], optional): Object rotation range.
zhangwenwei's avatar
zhangwenwei committed
223
224
225
226
            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
227
228

    def __init__(self,
zhangwenwei's avatar
zhangwenwei committed
229
                 translation_std=[0.25, 0.25, 0.25],
zhangwenwei's avatar
zhangwenwei committed
230
                 global_rot_range=[0.0, 0.0],
zhangwenwei's avatar
zhangwenwei committed
231
                 rot_range=[-0.15707963267, 0.15707963267],
zhangwenwei's avatar
zhangwenwei committed
232
                 num_try=100):
zhangwenwei's avatar
zhangwenwei committed
233
        self.translation_std = translation_std
zhangwenwei's avatar
zhangwenwei committed
234
        self.global_rot_range = global_rot_range
zhangwenwei's avatar
zhangwenwei committed
235
        self.rot_range = rot_range
zhangwenwei's avatar
zhangwenwei committed
236
237
238
        self.num_try = num_try

    def __call__(self, input_dict):
239
240
241
242
243
244
245
246
247
        """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
248
249
        gt_bboxes_3d = input_dict['gt_bboxes_3d']
        points = input_dict['points']
zhangwenwei's avatar
zhangwenwei committed
250

zhangwenwei's avatar
zhangwenwei committed
251
        # TODO: check this inplace function
252
        numpy_box = gt_bboxes_3d.tensor.numpy()
zhangwenwei's avatar
zhangwenwei committed
253
        noise_per_object_v3_(
254
            numpy_box,
zhangwenwei's avatar
zhangwenwei committed
255
            points,
zhangwenwei's avatar
zhangwenwei committed
256
257
            rotation_perturb=self.rot_range,
            center_noise_std=self.translation_std,
zhangwenwei's avatar
zhangwenwei committed
258
259
            global_random_rot_range=self.global_rot_range,
            num_try=self.num_try)
260
261

        input_dict['gt_bboxes_3d'] = gt_bboxes_3d.new_box(numpy_box)
zhangwenwei's avatar
zhangwenwei committed
262
263
264
265
        input_dict['points'] = points
        return input_dict

    def __repr__(self):
266
        """str: Return a string that describes the module."""
zhangwenwei's avatar
zhangwenwei committed
267
268
        repr_str = self.__class__.__name__
        repr_str += '(num_try={},'.format(self.num_try)
zhangwenwei's avatar
zhangwenwei committed
269
        repr_str += ' translation_std={},'.format(self.translation_std)
zhangwenwei's avatar
zhangwenwei committed
270
        repr_str += ' global_rot_range={},'.format(self.global_rot_range)
zhangwenwei's avatar
zhangwenwei committed
271
        repr_str += ' rot_range={})'.format(self.rot_range)
zhangwenwei's avatar
zhangwenwei committed
272
273
274
        return repr_str


275
@PIPELINES.register_module()
zhangwenwei's avatar
zhangwenwei committed
276
class GlobalRotScaleTrans(object):
zhangwenwei's avatar
zhangwenwei committed
277
    """Apply global rotation, scaling and translation to a 3D scene.
zhangwenwei's avatar
zhangwenwei committed
278
279
280

    Args:
        rot_range (list[float]): Range of rotation angle.
liyinhao's avatar
liyinhao committed
281
            Defaults to [-0.78539816, 0.78539816] (close to [-pi/4, pi/4]).
zhangwenwei's avatar
zhangwenwei committed
282
        scale_ratio_range (list[float]): Range of scale ratio.
liyinhao's avatar
liyinhao committed
283
            Defaults to [0.95, 1.05].
zhangwenwei's avatar
zhangwenwei committed
284
285
286
        translation_std (list[float]): The standard deviation of ranslation
            noise. This apply random translation to a scene by a noise, which
            is sampled from a gaussian distribution whose standard deviation
liyinhao's avatar
liyinhao committed
287
288
            is set by ``translation_std``. Defaults to [0, 0, 0]
        shift_height (bool): Whether to shift height.
wuyuefeng's avatar
wuyuefeng committed
289
            (the fourth dimension of indoor points) when scaling.
liyinhao's avatar
liyinhao committed
290
            Defaults to False.
zhangwenwei's avatar
zhangwenwei committed
291
    """
zhangwenwei's avatar
zhangwenwei committed
292
293

    def __init__(self,
zhangwenwei's avatar
zhangwenwei committed
294
295
                 rot_range=[-0.78539816, 0.78539816],
                 scale_ratio_range=[0.95, 1.05],
wuyuefeng's avatar
wuyuefeng committed
296
297
                 translation_std=[0, 0, 0],
                 shift_height=False):
zhangwenwei's avatar
zhangwenwei committed
298
299
300
        self.rot_range = rot_range
        self.scale_ratio_range = scale_ratio_range
        self.translation_std = translation_std
wuyuefeng's avatar
wuyuefeng committed
301
        self.shift_height = shift_height
zhangwenwei's avatar
zhangwenwei committed
302
303

    def _trans_bbox_points(self, input_dict):
304
305
306
307
308
309
310
311
312
313
        """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.
        """
zhangwenwei's avatar
zhangwenwei committed
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
        if not isinstance(self.translation_std, (list, tuple, np.ndarray)):
            translation_std = [
                self.translation_std, self.translation_std,
                self.translation_std
            ]
        else:
            translation_std = self.translation_std
        translation_std = np.array(translation_std, dtype=np.float32)
        trans_factor = np.random.normal(scale=translation_std, size=3).T

        input_dict['points'][:, :3] += trans_factor
        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):
330
331
332
333
334
335
336
337
338
339
        """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
340
        rotation = self.rot_range
zhangwenwei's avatar
zhangwenwei committed
341
342
343
        if not isinstance(rotation, list):
            rotation = [-rotation, rotation]
        noise_rotation = np.random.uniform(rotation[0], rotation[1])
zhangwenwei's avatar
zhangwenwei committed
344
345

        for key in input_dict['bbox3d_fields']:
wuyuefeng's avatar
wuyuefeng committed
346
347
348
349
350
            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
351

zhangwenwei's avatar
zhangwenwei committed
352
    def _scale_bbox_points(self, input_dict):
353
354
355
356
357
358
359
360
361
        """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
362
363
        scale = input_dict['pcd_scale_factor']
        input_dict['points'][:, :3] *= scale
wuyuefeng's avatar
wuyuefeng committed
364
365
366
        if self.shift_height:
            input_dict['points'][:, -1] *= scale

zhangwenwei's avatar
zhangwenwei committed
367
368
        for key in input_dict['bbox3d_fields']:
            input_dict[key].scale(scale)
zhangwenwei's avatar
zhangwenwei committed
369

zhangwenwei's avatar
zhangwenwei committed
370
    def _random_scale(self, input_dict):
371
372
373
374
375
376
377
378
379
        """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
380
381
382
        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
383
384

    def __call__(self, input_dict):
385
386
387
388
389
390
391
392
393
394
395
        """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.
        """
zhangwenwei's avatar
zhangwenwei committed
396
        self._rot_bbox_points(input_dict)
zhangwenwei's avatar
zhangwenwei committed
397

zhangwenwei's avatar
zhangwenwei committed
398
399
400
        if 'pcd_scale_factor' not in input_dict:
            self._random_scale(input_dict)
        self._scale_bbox_points(input_dict)
zhangwenwei's avatar
zhangwenwei committed
401

zhangwenwei's avatar
zhangwenwei committed
402
        self._trans_bbox_points(input_dict)
zhangwenwei's avatar
zhangwenwei committed
403
404
405
        return input_dict

    def __repr__(self):
406
        """str: Return a string that describes the module."""
zhangwenwei's avatar
zhangwenwei committed
407
        repr_str = self.__class__.__name__
zhangwenwei's avatar
zhangwenwei committed
408
409
410
        repr_str += '(rot_range={},'.format(self.rot_range)
        repr_str += ' scale_ratio_range={},'.format(self.scale_ratio_range)
        repr_str += ' translation_std={})'.format(self.translation_std)
wuyuefeng's avatar
wuyuefeng committed
411
        repr_str += ' shift_height={})'.format(self.shift_height)
zhangwenwei's avatar
zhangwenwei committed
412
413
414
        return repr_str


415
@PIPELINES.register_module()
zhangwenwei's avatar
zhangwenwei committed
416
class PointShuffle(object):
417
    """Shuffle input points."""
zhangwenwei's avatar
zhangwenwei committed
418
419

    def __call__(self, input_dict):
420
421
422
423
424
425
426
427
428
        """Call function to shuffle points.

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

        Returns:
            dict: Results after filtering, 'points' keys are updated \
                in the result dict.
        """
zhangwenwei's avatar
zhangwenwei committed
429
430
431
432
433
434
435
        np.random.shuffle(input_dict['points'])
        return input_dict

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


436
@PIPELINES.register_module()
zhangwenwei's avatar
zhangwenwei committed
437
class ObjectRangeFilter(object):
438
439
440
441
442
    """Filter objects by the range.

    Args:
        point_cloud_range (list[float]): Point cloud range.
    """
zhangwenwei's avatar
zhangwenwei committed
443
444
445
446
447
448

    def __init__(self, point_cloud_range):
        self.pcd_range = np.array(point_cloud_range, dtype=np.float32)
        self.bev_range = self.pcd_range[[0, 1, 3, 4]]

    def __call__(self, input_dict):
449
450
451
452
453
454
455
456
457
        """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.
        """
zhangwenwei's avatar
zhangwenwei committed
458
        gt_bboxes_3d = input_dict['gt_bboxes_3d']
zhangwenwei's avatar
zhangwenwei committed
459
        gt_labels_3d = input_dict['gt_labels_3d']
460
        mask = gt_bboxes_3d.in_range_bev(self.bev_range)
zhangwenwei's avatar
zhangwenwei committed
461
        gt_bboxes_3d = gt_bboxes_3d[mask]
ZwwWayne's avatar
ZwwWayne committed
462
463
464
465
466
        # 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
467
468

        # limit rad to [-pi, pi]
469
470
        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
471
472
        input_dict['gt_labels_3d'] = gt_labels_3d

zhangwenwei's avatar
zhangwenwei committed
473
474
475
        return input_dict

    def __repr__(self):
476
        """str: Return a string that describes the module."""
zhangwenwei's avatar
zhangwenwei committed
477
478
479
480
481
        repr_str = self.__class__.__name__
        repr_str += '(point_cloud_range={})'.format(self.pcd_range.tolist())
        return repr_str


482
@PIPELINES.register_module()
zhangwenwei's avatar
zhangwenwei committed
483
class PointsRangeFilter(object):
484
485
486
487
488
    """Filter points by the range.

    Args:
        point_cloud_range (list[float]): Point cloud range.
    """
zhangwenwei's avatar
zhangwenwei committed
489
490
491
492
493
494

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

    def __call__(self, input_dict):
495
496
497
498
499
500
501
502
503
        """Call function to filter points by the range.

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

        Returns:
            dict: Results after filtering, 'points' keys are updated \
                in the result dict.
        """
zhangwenwei's avatar
zhangwenwei committed
504
505
506
507
508
509
510
511
512
        points = input_dict['points']
        points_mask = ((points[:, :3] >= self.pcd_range[:, :3])
                       & (points[:, :3] < self.pcd_range[:, 3:]))
        points_mask = points_mask[:, 0] & points_mask[:, 1] & points_mask[:, 2]
        clean_points = points[points_mask, :]
        input_dict['points'] = clean_points
        return input_dict

    def __repr__(self):
513
        """str: Return a string that describes the module."""
zhangwenwei's avatar
zhangwenwei committed
514
515
516
        repr_str = self.__class__.__name__
        repr_str += '(point_cloud_range={})'.format(self.pcd_range.tolist())
        return repr_str
zhangwenwei's avatar
zhangwenwei committed
517
518
519
520


@PIPELINES.register_module()
class ObjectNameFilter(object):
zhangwenwei's avatar
zhangwenwei committed
521
    """Filter GT objects by their names.
zhangwenwei's avatar
zhangwenwei committed
522
523

    Args:
liyinhao's avatar
liyinhao committed
524
        classes (list[str]): List of class names to be kept for training.
zhangwenwei's avatar
zhangwenwei committed
525
526
527
528
529
530
531
    """

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

    def __call__(self, input_dict):
532
533
534
535
536
537
538
539
540
        """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
541
542
543
544
545
546
547
548
549
        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):
550
        """str: Return a string that describes the module."""
zhangwenwei's avatar
zhangwenwei committed
551
552
553
        repr_str = self.__class__.__name__
        repr_str += f'(classes={self.classes})'
        return repr_str
wuyuefeng's avatar
wuyuefeng committed
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579


@PIPELINES.register_module()
class IndoorPointSample(object):
    """Indoor point sample.

    Sampling data to a certain number.

    Args:
        name (str): Name of the dataset.
        num_points (int): Number of points to be sampled.
    """

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

    def points_random_sampling(self,
                               points,
                               num_samples,
                               replace=None,
                               return_choices=False):
        """Points random sampling.

        Sample points to a certain number.

        Args:
580
            points (np.ndarray): 3D Points.
wuyuefeng's avatar
wuyuefeng committed
581
582
            num_samples (int): Number of samples to be sampled.
            replace (bool): Whether the sample is with or without replacement.
liyinhao's avatar
liyinhao committed
583
584
            Defaults to None.
            return_choices (bool): Whether return choice. Defaults to False.
wuyuefeng's avatar
wuyuefeng committed
585
586

        Returns:
587
588
589
590
            tuple[np.ndarray] | np.ndarray:

                - points (np.ndarray): 3D Points.
                - choices (np.ndarray, optional): The generated random samples.
wuyuefeng's avatar
wuyuefeng committed
591
592
593
594
595
596
597
598
599
600
601
        """
        if replace is None:
            replace = (points.shape[0] < num_samples)
        choices = np.random.choice(
            points.shape[0], num_samples, replace=replace)
        if return_choices:
            return points[choices], choices
        else:
            return points[choices]

    def __call__(self, results):
602
603
604
605
606
607
608
609
610
        """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
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
        points = results['points']
        points, choices = self.points_random_sampling(
            points, self.num_points, return_choices=True)
        pts_instance_mask = results.get('pts_instance_mask', None)
        pts_semantic_mask = results.get('pts_semantic_mask', None)
        results['points'] = points

        if pts_instance_mask is not None and pts_semantic_mask is not None:
            pts_instance_mask = pts_instance_mask[choices]
            pts_semantic_mask = pts_semantic_mask[choices]
            results['pts_instance_mask'] = pts_instance_mask
            results['pts_semantic_mask'] = pts_semantic_mask

        return results

    def __repr__(self):
627
        """str: Return a string that describes the module."""
wuyuefeng's avatar
wuyuefeng committed
628
629
630
        repr_str = self.__class__.__name__
        repr_str += '(num_points={})'.format(self.num_points)
        return repr_str