作者: Akinori Nagano , Masahiro Fujimoto , Shoma Kudo , Ryosuke Akaguma
DOI: 10.1016/J.GAITPOST.2016.09.015
关键词: Artificial intelligence 、 High-speed camera 、 Computer vision 、 Coordinate system 、 Root mean square 、 Image processing 、 Motion capture 、 Centroid 、 STRIDE 、 Position (vector) 、 Computer science
摘要: Abstract Walking and running speed is a fundamental parameter studied in wide range of areas such as sport biomechanics, rehabilitation, health promotion the elderly, etc. Given that walking or not constant even within stride, instantaneous changes body motion need to be evaluated better understand one’s performance. In this study, new cost- time- efficient methodology determine horizontal was developed. The newly developed method processes movies taken with (high-speed) camera. It consists five sub-steps, which are performed serial order: (1) Subtraction background image, (2) filtering, (3) binarization centroid determination, (4) transformation laboratory coordinate system (5) differentiation. To test accuracy method, output (position speed) compared data obtained using capture. average root mean squared (RMS) error (difference between outputs capture) position-time curves 0.011 m–0.033 m. RMS speed-time 0.054 m/s–0.076 m/s. shown produces accurate speed.