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
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
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
| """
基于 3D ToF 的 OOP 检测
使用 IDS Nion 深度图检测异常乘员姿态
"""
import numpy as np
from typing import Dict, List, Tuple
class OOPDetector:
"""
Out-of-Position 异常姿态检测器
使用 3D ToF 深度图
"""
def __init__(self, config: dict = None):
self.seat_positions = {
'driver': {'center': (200, 400, 500), 'size': (400, 400, 600)},
'passenger': {'center': (600, 400, 500), 'size': (400, 400, 600)},
'rear_left': {'center': (200, 800, 500), 'size': (400, 400, 600)},
'rear_right': {'center': (600, 800, 500), 'size': (400, 400, 600)},
}
self.oop_thresholds = {
'head_to_dashboard': 300,
'chest_to_airbag': 250,
'feet_on_dashboard': True,
'leaning_forward': 200,
'reclined_excessively': 45,
}
def detect_oop(
self,
depth_map: np.ndarray,
camera_params: dict
) -> Dict:
"""
检测异常姿态
Args:
depth_map: 深度图 (H, W)，单位 mm
camera_params: 相机内参 {
'fx': float,
'fy': float,
'cx': float,
'cy': float
}
Returns:
{
'is_oop': bool,
'oop_type': str,
'severity': str,
'affected_zone': str
}
"""
point_cloud = self.depth_to_pointcloud(depth_map, camera_params)
occupant_regions = self.segment_occupants(point_cloud)
oop_results = []
for seat_name, points in occupant_regions.items():
pose = self.estimate_pose(points)
oop = self.check_oop_conditions(pose, seat_name)
if oop['is_oop']:
oop_results.append({
'seat': seat_name,
**oop
})
return oop_results
def depth_to_pointcloud(
self,
depth_map: np.ndarray,
params: dict
) -> np.ndarray:
"""
深度图转 3D 点云
Args:
depth_map: (H, W) 深度图，单位 mm
params: 相机内参
Returns:
points: (N, 3) 点云，单位 mm
"""
H, W = depth_map.shape
u = np.arange(W)
v = np.arange(H)
u, v = np.meshgrid(u, v)
fx, fy = params['fx'], params['fy']
cx, cy = params['cx'], params['cy']
x = (u - cx) * depth_map / fx
y = (v - cy) * depth_map / fy
z = depth_map
valid = depth_map > 0
points = np.stack([x[valid], y[valid], z[valid]], axis=-1)
return points
def estimate_pose(self, points: np.ndarray) -> Dict:
"""
从点云估计 3D 姿态
Returns:
{
'head_position': (x, y, z),
'chest_position': (x, y, z),
'orientation': (roll, pitch, yaw),
'skeleton': {...}
}
"""
from sklearn.cluster import DBSCAN
clustering = DBSCAN(eps=50, min_samples=100).fit(points)
labels = clustering.labels_
unique_labels, counts = np.unique(labels[labels >= 0], return_counts=True)
main_label = unique_labels[np.argmax(counts)]
main_points = points[labels == main_label]
min_coords = main_points.min(axis=0)
max_coords = main_points.max(axis=0)
head_y_threshold = max_coords[1] - (max_coords[1] - min_coords[1]) * 0.3
head_points = main_points[main_points[:, 1] < head_y_threshold]
chest_points = main_points[main_points[:, 1] >= head_y_threshold]
head_position = head_points.mean(axis=0) if len(head_points) > 0 else None
chest_position = chest_points.mean(axis=0) if len(chest_points) > 0 else None
return {
'head_position': head_position,
'chest_position': chest_position,
'bounding_box': (min_coords, max_coords),
'point_count': len(main_points)
}
def check_oop_conditions(
self,
pose: Dict,
seat_name: str
) -> Dict:
"""
检查 OOP 条件
"""
oop_violations = []
head = pose['head_position']
chest = pose['chest_position']
bbox_min, bbox_max = pose['bounding_box']
seat = self.seat_positions.get(seat_name)
if not seat:
return {'is_oop': False}
if head is not None and seat_name == 'driver':
dashboard_x = seat['center'][0] - seat['size'][0] / 2 - 100
head_to_dashboard = head[0] - dashboard_x
if head_to_dashboard < self.oop_thresholds['head_to_dashboard']:
oop_violations.append('head_too_close')
if chest is not None:
airbag_x = seat['center'][0] - seat['size'][0] / 2
chest_to_airbag = chest[0] - airbag_x
if chest_to_airbag < self.oop_thresholds['chest_to_airbag']:
oop_violations.append('chest_too_close')
if head is not None and chest is not None:
forward_lean = seat['center'][0] - head[0]
if forward_lean > self.oop_thresholds['leaning_forward']:
oop_violations.append('leaning_forward')
if len(oop_violations) >= 2:
severity = 'severe'
elif len(oop_violations) == 1:
severity = 'moderate'
else:
severity = 'none'
return {
'is_oop': len(oop_violations) > 0,
'oop_type': oop_violations,
'severity': severity
}
if __name__ == "__main__":
camera_params = {
'fx': 1000,
'fy': 1000,
'cx': 640,
'cy': 480
}
depth_map = np.random.randint(500, 2000, (960, 1280))
detector = OOPDetector()
results = detector.detect_oop(depth_map, camera_params)
print(f"检测到 {len(results)} 个 OOP")
for r in results:
print(f" 座位: {r['seat']}, 类型: {r['oop_type']}, 严重程度: {r['severity']}")
|