以慣性感測與攝影影像為基礎之跌倒偵測

Fall Detection Based on Inertial Sensors and Video Camera

從預防跌倒、跌倒偵測及跌到後照護是老年人健康維護之重要議題，尤其現今穿戴式科技及攝影監視的普及，跌倒偵測更是近年研發焦點。本文針對以慣性感測或以攝影影像為基礎之跌倒偵測進行文獻回顧；並介紹本研究團隊近年所開發之以身體加速度偵測跌倒，及以人體輪廓影像模型特徵變化偵測跌倒之方法；最後提出跌倒偵測議題之未來展望。

Falling including prediction, detection, and post-fall care are important issues in health maintenance of the elderly. In particular, the widespread applications of wearable devices and video camera surveillance systems appeal much more attentions in fall detection. In this paper, a review of fall detections based on inertial sensors and video camera are presented. Meanwhile, novel fall detection methods based on body accelerations or human body contour model from video image are introduced. Finally, future issues about fall detection are addressed.