Evaluation System of Push-up Action Based on Kinect

In order to realize automatic counting and motion evaluation of push-ups, Kinect sensor is used to collect obtain skeleton key points. By analyzing the rules and characteristics of push-up movement, the required key points are selected. The effective data are obtained by data processing methods such as vacancy filling, data conversion and mean filtering. By extracting the characteristics of push-ups, segmentation is carried out to realize movement counting. DTW algorithm is used to calculate the relative approximation of each push-up movement and the pre-set standard movement, so as to measure whether the movement is standard. The test shows that the system is feasible.