為交通燈裝上「眼睛」：SingTown科技的人工智能視覺輔助

In urban traffic management, accurately identifying pedestrians within crosswalks and promptly triggering safety warnings is a crucial requirement for intelligent transportation systems.

傳統雷達或紅外線探測在複雜環境中難以準確區分行人與車輛，導致誤報率高且可靠性不足。

為此，交通智能設備製造商採用SingTown的OpenMV智能相機識別技術，利用AI機器視覺算法及區域檢測模型，實現對指定區域（如斑馬線）的動態監控及行人檢測預警。

Traffic intelligent equipment manufacturers utilise SingTown's OpenMV intelligent camera recognition technology to achieve real-time pedestrian detection at crosswalks and signal linkage warnings.

The intelligent vision module is installed at road intersections, delineating a rectangular monitoring area and detecting only moving targets within this region.

When a moving pedestrian target is detected, the camera immediately outputs a high-level signal via GPIO, which can be linked to signal lights, buzzers, or traffic control systems for alarm notification.

The system leverages the edge deployment advantages of embedded vision and on-device AI, enabling real-time pedestrian recognition through background modelling, target motion detection, and morphological recognition algorithms. It performs image analysis and signal output locally without requiring additional servers, with a recognition latency of less than 200 milliseconds.

The system can adapt to various lighting conditions and weather environments, featuring strong robustness and low power consumption characteristics.

透過SingTown Technology嘅OpenMV智能相機識別技術，交通智能設備製造商實現咗由傳統感應檢測到AI視覺行人識別同預警嘅智能升級，為城市交通安全提供高效實時嘅視覺解決方案。