Human Gait Analysis Method Based on Kinect Sensor

摘要

In this paper, a new method for human gait analysis based on the Kinect Sensor is introduced. Such method based on Kinect sensor can be divided into three steps: data acquisition, pre-processing and gait parameter calculation. First, a GUI (Graphical User Interface) was designed to control the Kinect sensor and get the required raw gait data. In the pre-processing, abnormal frames are removed first. Afterwards, the influence of Kinect's installation error is eliminated by coordinate system transformation. What's more, the noise is eliminated by using moving average filtering and median filtering. Finally, gait parameters are obtained by the designed algorithm which composed of gait cycle detection, gait parameter calculation, and gait phase extraction. The validity of the gait analysis method based on Kinect v2 was verified by experiments.