-
Notifications
You must be signed in to change notification settings - Fork 12
/
utils.py
executable file
·75 lines (57 loc) · 2.11 KB
/
utils.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
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
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
import numpy as np
def calc_depth_grad(depth):
'''
Calculate depth gradient
:param depth: depth map
:return: depth gradient
'''
# [1 0 -1] kernel
depth_grad_y = np.abs(depth[2:] - depth[:-2])
depth_grad_x = np.abs(depth[:, 2:] - depth[:, :-2])
depth_grad = np.zeros_like(depth)
depth_grad[1:-1] = depth_grad_y
depth_grad[:, 1:-1] = np.maximum(depth_grad[:, 1:-1], depth_grad_x)
return depth_grad
def get_coord_grid(h, w):
'''
Get coordinate grid
:param h: image height
:param w: image width
:return: coordinate grid (3, h * w), coordinate grid (h, w, 3)
'''
y_coords, x_coords = np.meshgrid(range(h), range(w), indexing='ij')
z_coords = np.ones_like(x_coords)
coordinate_grid_flat = np.array([x_coords, y_coords, z_coords])
coordinate_grid_flat = np.reshape(coordinate_grid_flat, newshape=(3, -1))
coordinate_grid_2d = np.concatenate((x_coords[:, :, None], y_coords[:, :, None], z_coords[:, :, None]), axis=2)
return coordinate_grid_flat, coordinate_grid_2d
def calc_plane_type(plane_mask_flat, plane_seg_flat_sum_classes, val_ind):
'''
Calculate plane type
:param plane_mask_flat: Plane mask
:param plane_seg_flat_sum_classes: Number of points on plane for each of the segmentation classes
:param val_ind: valid segmentation indices
:return: Plane type
'''
plane_type = 0
plane_type_area = 0
for plane_type_cand in val_ind:
plane_type_cand_area = np.sum(plane_seg_flat_sum_classes[plane_type_cand]) / np.sum(plane_mask_flat)
if plane_type_cand_area > plane_type_area:
plane_type_area = plane_type_cand_area
plane_type = plane_type_cand
if plane_type > 0:
return plane_type
else:
return -1
def norm_plane_params(plane_params):
'''
Normalize plane parameters
:param plane_params:
:return:
'''
plane_params = plane_params / np.sqrt(
plane_params[0] ** 2 + plane_params[1] ** 2 + plane_params[2] ** 2)
if plane_params.shape[0] == 4 and plane_params[-1] < 0.:
plane_params *= -1
return plane_params