计算机视觉视频图像处理在暖通空调控制信号采集领域的应用
Application of computer vision/video image processing in collecting HVAC control signals
摘要：
以人为本的智能建筑设计运行，通过给建筑装上“眼睛”，达到实时监测室内人员相关信息的目的，比如人员的定位、在室情况、冷热状态等。近年来，人工智能领域的计算机视觉视频图像处理技术的发展突飞猛进，为上述信号的实时非接触监测提供了保障。基于近5~10年的相关研究成果，系统梳理了计算机视觉视频图像处理技术在建筑内人员行为、冷热姿态、生理参数测量等方面的研究成果。分析了红外探测器、基于欧拉视频放大技术的普通摄像头探测器、人体骨骼关键点模型等技术的特点、应用及未来发展方向，同时展望了其在按需通风、个体环境控制、睡眠环境监测等领域的应用前景。
Abstract：
People-oriented design and operation are widely used in smart buildings. In the control loop, computer vision technologies (building “eyes”) can achieve the purpose of real time monitoring and identifying occupant related information, such as personnel positioning, occupied/unoccupied spaces, hot/cold conditions. In recent years, computer vision/video image processing technologies have been developed rapidly, which paves the way for real time non-contact monitoring. Based on relevant research achievements in the past five to ten years, systematically reviews research results and their robustness (pros and cons) of computer vision/video image processing technologies in identifying occupant’s behavior, catching hot/cold poses, monitoring physiological parameters, etc. Analyses the characteristics, applications and future development directions of infrared detectors, ordinary camera detectors based on Euler’s video amplification technology and human bone key point models. Envisages their application prospects in demand-controlled ventilation, individual environment control, and sleep environment monitoring.
Keywords:computervision,videoimageprocessing,occupantbehavior,hot/coldpose,physiologicalparameter,smartbuilding