Action semantic information identification method based on data enhancement and intelligent motion segmentation

Abstract

The invention relates to an action semantic information identification method and system based on data enhancement and intelligent motion segmentation. The method comprises the following steps: segmenting six-axis inertial sensor motion data by adopting an intelligent motion segmentation algorithm based on waveform-track information fusion to obtain a plurality of six-axis inertial sensor data fragments; carrying out data enhancement processing on the six-axis inertial sensor data segment based on a signal reconstruction algorithm to generate a six-axis inertial sensor waveform data segment after data enhancement, and carrying out trajectory calculation based on the six-axis inertial sensor waveform data segment after data enhancement to generate a three-dimensional motion trajectory corresponding to the six-axis inertial sensor waveform data segment after data enhancement; and performing up-sampling processing on the waveform data fragment of the six-axis inertial sensor after data enhancement and the three-dimensional motion trail of the waveform data fragment, and inputting the waveform data fragment after up-sampling and the three-dimensional motion trail of the waveform data fragment into a space graph neural network for semantic information recognition to obtain a recognition result of action semantic information. And high-precision action semantic information identification based on the inertial sensor is realized.

Description

technical field [0001] The invention relates to the technical field of human-computer interaction control based on inertial sensors, in particular to a method and system for recognizing action semantic information based on data enhancement and intelligent motion segmentation. Background technique [0002] MEMS (Micro-Electro-Mechanical System) inertial sensors have the advantages of small size, easy wearing, low power consumption, low cost, and easy mass production. The unit price of common inertial sensors on the market can be as low as 0.2-0.5 yuan. Therefore, based on The human-computer interaction of inertial sensors is not only more natural (compared with voice, image, etc., it is not limited by distance and scene), but also requires less hardware configuration, so it has a huge market prospect. However, the accuracy of inertial sensors is poor, especially the errors of cheap inertial sensors are difficult to support high-precision attitude calculation, trajectory resto...

AI Technical Summary

Problems solved by technology

[0006] At present, there are some mature Human Activity Recognition (HAR) schemes based on inertial sensors, such as detecting walking, running, jumping, going up and down stairs, etc. However, these schemes are difficult to apply to semantic information detection of short-term motion. Therefore, human-computer interaction control cannot be realized

For the short-term action semantic information recognition task, the acquisition of short-term action corresponding data depends on accurate motion segmentation technology, and this field is not yet mature

Images