"""
Glasgow AI 路侧摄像头系统技术分析
基于公开报道推断的技术架构

关键信息：
- 部署位置：路边固定杆
- 检测目标：手机使用 + 安全带违规
- 技术路线：AI视觉 + 多帧验证
- 来源：Traffic Scotland Seatbelt and Driver Distraction Survey
"""

import numpy as np

class RoadsideAICamera:
    """路侧 AI 摄像头系统"""
    
    def __init__(self):
        self.detection_range = 15  # 米
        self.camera_height = 5.5  # 米
        self.mounting_type = "roadside_pole"
        self.fov_h = 40  # 度，窄视角
        self.fov_v = 30
        self.resolution = (4096, 2160)  # 4K
        
        # 检测模型
        self.models = {
            "phone_detection": {
                "model": "YOLOv8-m",
                "classes": ["phone_hand", "phone_ear"],
                "min_confidence": 0.85,
                "frames_needed": 3,  # 至少3帧确认
            },
            "seatbelt_detection": {
                "model": "ResNet-50",
                "classes": ["belt_on", "belt_off", "belt_misuse"],
                "min_confidence": 0.90,
                "frames_needed": 5,
            },
        }
        
    def analyze_detection_accuracy(self, vehicle_speed: float) -> dict:
        """
        分析不同车速下的检测能力
        
        Args:
            vehicle_speed: 车速 km/h
        """
        # 车辆通过检测区域时间
        detection_window = (self.detection_range * 2) / (vehicle_speed / 3.6)
        
        # 30fps 摄像头可捕获帧数
        frames_captured = int(detection_window * 30)
        
        # 确认所需帧数
        phone_frames_needed = self.models["phone_detection"]["frames_needed"]
        belt_frames_needed = self.models["seatbelt_detection"]["frames_needed"]
        
        phone_feasible = frames_captured >= phone_frames_needed
        belt_feasible = frames_captured >= belt_frames_needed
        
        # 分辨率需求
        # 手机: 10cm @ 15m → 需要~4K
        # 安全带: 5cm @ 15m → 需要~8K
        
        return {
            "speed_kmh": vehicle_speed,
            "detection_window_s": round(detection_window, 2),
            "frames_captured": frames_captured,
            "phone_detection": "✅" if phone_feasible else "❌",
            "seatbelt_detection": "✅" if belt_feasible else "❌",
            "resolution_sufficient": "phone:✅ belt:⚠️",
        }

camera = RoadsideAICamera()

print("=== Glasgow AI 摄像头检测能力分析 ===")
print(f"{'车速(km/h)':<15} {'检测时间(s)':<15} {'捕获帧数':<10} {'手机':<10} {'安全带':<10}")
print("-" * 60)
for speed in [30, 50, 70, 90, 110, 130]:
    r = camera.analyze_detection_accuracy(speed)
    print(f"{r['speed_kmh']:<15} {r['detection_window_s']:<15} {r['frames_captured']:<10} {r['phone_detection']:<10} {r['seatbelt_detection']:<10}")

print("\n⚠️ 高速时检测窗口不足，需多摄像头接力或更高帧率")

"""
西澳大利亚 Great Southern 地区 AI 摄像头效果

关键数据（2025.6 - 2026.7）：
- 总体违规率下降 82%
- 手机违规下降 85%
- 超速违规下降 83%

来源：soPerth.com.au 报道
"""

class AICameraEffectiveness:
    """AI 摄像头效果分析"""
    
    def __init__(self):
        self.data = {
            "period": "2025.6 - 2026.7 (13个月)",
            "location": "Great Southern, WA",
            "metrics": {
                "overall_offence_reduction": 82,
                "mobile_phone_reduction": 85,
                "speeding_reduction": 83,
            }
        }
        
    def analyze(self):
        print("=== 西澳 AI 摄像头 13 个月效果 ===")
        for metric, reduction in self.data["metrics"].items():
            print(f"{metric}: -{reduction}%")
        
        print(f"\n⚠️ 关键洞察:")
        print(f"1. 威慑效应 >> 检测效应（大多数驾驶员因知道有摄像头而守规）")
        print(f"2. 手机检测效果最好（-85%），说明视觉可检测场景明确")
        print(f"3. 车内DMS互补价值：车内可在无路侧摄像头的道路提供检测")

effect = AICameraEffectiveness()
effect.analyze()

"""
安全带检测：外部 vs 内部技术路线对比
"""

class SeatbeltDetectionComparison:
    """安全带检测方案对比"""
    
    COMPARISON = {
        "路侧AI摄像头": {
            "method": "4K高清 + ResNet 分类",
            "accuracy": "~90%",
            "limitation": "挡风玻璃反光、车速>100km/h窗口不足",
            "cost": "$15K/点位/年",
            "coverage": "仅固定路线",
        },
        "车内DMS摄像头": {
            "method": "IR摄像头 + 关键点检测",
            "accuracy": "~95%",
            "limitation": "深色安全带+深色衣物难区分",
            "cost": "<$5/车",
            "coverage": "全时全路",
        },
        "安全带张力传感器": {
            "method": "机械张力检测",
            "accuracy": "~99%",
            "limitation": "只能检测是否系上，不能检测位置",
            "cost": "~$10/车",
            "coverage": "全时全路",
        },
        "多模态融合": {
            "method": "DMS摄像头 + 张力传感器 + 视觉位置检测",
            "accuracy": "~98%",
            "limitation": "成本增加",
            "cost": "~$15/车",
            "coverage": "全时全路 + 位置检测",
        },
    }
    
    def print_comparison(self):
        for method, info in self.COMPARISON.items():
            print(f"\n{method}:")
            for key, val in info.items():
                print(f"  {key}: {val}")

comp = SeatbeltDetectionComparison()
comp.print_comparison()