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
235
236
237
238
239
240
241
242
243
244
| """
Cineon+Smart Eye MicroBird 航空座舱眼动追踪
跨域对比: 航空 vs 汽车DMS
核心启示: 航空训练的"仪表扫描评估"可迁移到驾驶员分心检测
"""
import numpy as np
from typing import List, Dict, Tuple
class MicroBirdEyeTracking:
"""
Smart Eye MicroBird: 环境眼动追踪系统
与汽车DMS的差异:
1. 多摄像头覆盖整个座舱(非单摄像头)
2. 高精度仪表级映射(非粗粒度区域)
3. 训练评估而非实时警告
"""
def __init__(self):
self.cameras = {
'overhead': {'pos': [0, -0.3, 0.5], 'fov': 60},
'left_panel': {'pos': [-0.4, 0, 0.3], 'fov': 45},
'right_panel': {'pos': [0.4, 0, 0.3], 'fov': 45},
'center_dash': {'pos': [0, 0.1, 0.4], 'fov': 50},
}
self.instrument_aois = {
'ASI': {'center': (0.5, 0.3), 'size': (0.08, 0.08)},
'ALT': {'center': (0.6, 0.3), 'size': (0.08, 0.08)},
'HIS': {'center': (0.5, 0.4), 'size': (0.08, 0.08)},
'VSI': {'center': (0.6, 0.4), 'size': (0.08, 0.08)},
'PFD': {'center': (0.55, 0.35), 'size': (0.15, 0.15)},
'MFD': {'center': (0.75, 0.35), 'size': (0.12, 0.12)},
'throttle': {'center': (0.3, 0.7), 'size': (0.08, 0.08)},
'outside': {'center': (0.5, 0.1), 'size': (0.4, 0.15)},
}
def map_gaze_to_instrument(self, gaze_x, gaze_y) -> str:
"""将注视点映射到具体仪表"""
for instrument, aoi in self.instrument_aois.items():
cx, cy = aoi['center']
w, h = aoi['size']
if (abs(gaze_x - cx) < w/2 and
abs(gaze_y - cy) < h/2):
return instrument
return 'unknown'
def analyze_scan_pattern(self, gaze_sequence) -> Dict:
"""
分析仪表扫描模式
航空训练关键指标:
1. 仪表驻留时间 (dwell time)
2. 扫视频率 (saccade rate)
3. 仪表覆盖 (instrument coverage)
4. 外视时间 (outside dwell)
5. 扫描规律性 (scan regularity)
"""
timestamps = [g[0] for g in gaze_sequence]
instruments = [self.map_gaze_to_instrument(g[1], g[2])
for g in gaze_sequence]
dwell_times = {}
for i, inst in enumerate(instruments):
if i > 0:
dt = timestamps[i] - timestamps[i-1]
dwell_times[inst] = dwell_times.get(inst, 0) + dt
saccades = sum(1 for i in range(1, len(instruments))
if instruments[i] != instruments[i-1])
duration = timestamps[-1] - timestamps[0]
saccade_rate = saccades / duration if duration > 0 else 0
unique_instruments = set(instruments) - {'unknown'}
coverage = len(unique_instruments) / len(self.instrument_aois)
outside_time = dwell_times.get('outside', 0)
outside_ratio = outside_time / duration if duration > 0 else 0
inst_transitions = []
for i in range(1, len(instruments)):
if instruments[i] != instruments[i-1]:
inst_transitions.append(timestamps[i] - timestamps[i-1])
if len(inst_transitions) > 2:
regularity = np.std(inst_transitions) / np.mean(inst_transitions)
else:
regularity = float('inf')
return {
'dwell_times': dwell_times,
'saccade_rate': saccade_rate,
'coverage': coverage,
'outside_ratio': outside_ratio,
'scan_regularity': regularity,
'assessment': self._assess_scan(regularity, coverage, outside_ratio)
}
def _assess_scan(self, regularity, coverage, outside_ratio):
"""评估扫描质量"""
scores = []
if regularity < 0.5:
scores.append(('regularity', 'good', '扫描规律性好'))
elif regularity < 1.0:
scores.append(('regularity', 'fair', '扫描一般'))
else:
scores.append(('regularity', 'poor', '扫描不规律'))
if coverage > 0.7:
scores.append(('coverage', 'good', '仪表覆盖充分'))
else:
scores.append(('coverage', 'poor', '仪表覆盖不足'))
if 0.6 < outside_ratio < 0.8:
scores.append(('outside', 'good', '外视时间适当'))
elif outside_ratio < 0.6:
scores.append(('outside', 'poor', '外视不足(仪表盯着)'))
else:
scores.append(('outside', 'fair', '外视过多(忽略仪表)'))
return scores
class AviationToAutomotiveTransfer:
"""
航空眼动追踪 → 汽车DMS 跨域迁移
核心启示:
1. 仪表扫描分析 → 驾驶员分心检测
2. 情境感知评估 → 认知负荷估计
3. 多摄像头覆盖 → 全座舱感知
4. 处方镜兼容 → 更广人群覆盖
"""
@staticmethod
def transfer_scan_analysis():
"""航空扫描分析 → 汽车分心检测"""
transfers = {
'仪表驻留时间': {
'aviation': '各仪表驻留时间→训练评估',
'automotive': '仪表盘/后视镜/手机驻留→分心检测',
'ims_value': '量化分心程度'
},
'外视时间比': {
'aviation': '>60%看外面→安全飞行',
'automotive': '>80%看前方道路→安全驾驶',
'ims_value': '视线偏离道路时间'
},
'扫描规律性': {
'aviation': '规律扫描→训练有素',
'automotive': '规律扫描→正常驾驶',
'ims_value': '不规律=疲劳/分心'
},
'仪表覆盖率': {
'aviation': '覆盖所有关键仪表',
'automotive': '覆盖后视镜/侧镜/仪表盘',
'ims_value': '遗漏关键区域=认知缺失'
},
}
return transfers
if __name__ == "__main__":
tracker = MicroBirdEyeTracking()
print("=" * 60)
print("MicroBird 航空座舱眼动追踪 vs 汽车DMS")
print("=" * 60)
print(f"\n座舱摄像头配置:")
for name, cfg in tracker.cameras.items():
print(f" {name}: 位置{cfg['pos']}, FOV={cfg['fov']}°")
print(f"\n仪表AOI定义:")
for inst, aoi in tracker.instrument_aois.items():
print(f" {inst}: 中心{aoi['center']}, 大小{aoi['size']}")
np.random.seed(42)
n = 60 * 60
timestamps = np.arange(n) / 60
gaze_x = np.zeros(n)
gaze_y = np.zeros(n)
for i in range(n):
if np.random.rand() < 0.7:
gaze_x[i] = 0.5 + np.random.normal(0, 0.1)
gaze_y[i] = 0.1 + np.random.normal(0, 0.05)
else:
inst = np.random.choice(['ASI', 'ALT', 'HIS', 'VSI'])
aoi = tracker.instrument_aois[inst]
gaze_x[i] = aoi['center'][0] + np.random.normal(0, 0.02)
gaze_y[i] = aoi['center'][1] + np.random.normal(0, 0.02)
gaze_sequence = list(zip(timestamps, gaze_x, gaze_y))
analysis = tracker.analyze_scan_pattern(gaze_sequence)
print(f"\n扫描分析结果 (60秒模拟):")
print(f" 扫视频率: {analysis['saccade_rate']:.1f} 次/秒")
print(f" 仪表覆盖率: {analysis['coverage']:.1%}")
print(f" 外视时间比: {analysis['outside_ratio']:.1%}")
print(f" 扫描规律性(CV): {analysis['scan_regularity']:.2f}")
print(f"\n驻留时间分布:")
total_time = sum(analysis['dwell_times'].values())
for inst, t in sorted(analysis['dwell_times'].items(),
key=lambda x: -x[1]):
print(f" {inst}: {t:.1f}s ({t/total_time:.1%})")
print(f"\n评估结果:")
for dim, level, desc in analysis['assessment']:
print(f" {dim}: {level} - {desc}")
transfer = AviationToAutomotiveTransfer.transfer_scan_analysis()
print(f"\n{'='*60}")
print(f"航空 → 汽车 跨域迁移:")
print(f"{'指标':<15} {'航空应用':<25} {'汽车应用':<25} {'IMS价值'}")
print(f"{'-'*60}")
for metric, info in transfer.items():
print(f"{metric:<15} {info['aviation']:<25} {info['automotive']:<25} {info['ims_value']}")
|
