Department of Mining Engineering, Geomatics and Computer Graphics Research Group, Universidad de Oviedo, Mieres, Asturias, Spain.
Department of Manufacturing Engineering, Universidad de Oviedo, Gijón, Asturias, Spain.
PLoS One. 2018 Apr 11;13(4):e0195290. doi: 10.1371/journal.pone.0195290. eCollection 2018.
In cases of natural disasters, epidemics or even in dangerous situations like an act of terrorism, battle fields, a shooting or a mountain accident, finding survivors is a challenge. In these kind of situations it is sometimes critical to know if a person has vital signs or not, without the need to be in contact with the victim, thus avoiding jeopardizing the lives of the rescue workers. In this work, we propose the use of video magnification techniques to detect small movements in human bodies due to breathing that are invisible to the naked eye. Two different video magnification techniques, intensity-based and phase-based, were tested. The utility of these techniques to detect people who are alive but injured in risk situations was verified by simulating a scene with three people involved in an accident. Several factors such as camera stability, distance to the object, light conditions, magnification factor or computing time were analyzed. The results obtained were quite positive for both techniques, intensity-based method proving more adequate if the interest is in almost instant results whereas the phase-based method is more appropriate if processing time is not so relevant but the degree of magnification without excessive image noise.
在自然灾害、疫情，甚至在恐怖主义行为、战场、枪击或山体事故等危险情况下，寻找幸存者是一项挑战。在这些情况下，有时需要在无需接触受害者的情况下，确定一个人是否有生命体征，从而避免危及救援人员的生命。在这项工作中，我们提出使用视频放大技术来检测由于呼吸而导致的肉眼无法看到的人体微小运动。测试了两种不同的视频放大技术，基于强度和基于相位的技术。通过模拟涉及三人事故的场景，验证了这些技术在检测处于危险情况但受伤的人员方面的实用性。分析了摄像机稳定性、与物体的距离、光照条件、放大倍数或计算时间等多个因素。这两种技术的结果都非常积极，基于强度的方法如果对即时结果感兴趣，则更合适，而基于相位的方法如果处理时间不是那么重要，但放大倍数又没有过多的图像噪声，则更合适。
