CSI sensing identification method based on video analysis

Abstract

The invention discloses a CSI sensing identification method based on video analysis. The method comprises the following steps: step 1, collecting video data and CSI data; 2, screening the action data to obtain screened video data and CSI data; Step 3, counting the position and angle of an action sample for the screened video data, and supplementing the action sample according to a comparison result between the position and angle of the action sample and a threshold value to obtain a CSI data set; Step 4, preprocessing to obtain a signal feature segment corresponding to the action of the CSI data; and calculating the distance between the to-be-tested CSI data and each piece of CSI data in the CSI data set, wherein the category label corresponding to the minimum distance value is the category label of the to-be-tested data. According to the invention, video and CSI data acquisition are combined, the problem that a CSI data screening method is limited is solved to a certain extent, and the problem that the generalization ability of an identification model is poor due to the fact that action distribution is not wide enough and data diversity is poor during CSI data acquisition is effectively solved.

Description

technical field [0001] The invention belongs to the field of CSI-based perception recognition, and in particular relates to a video analysis-based CSI perception recognition method. Background technique [0002] The channel state information (CSI) measurement technology is to measure the channel frequency response on each subcarrier. The signal is sent from the transmitting end, after reflection and scattering of surrounding objects, and arrives at the receiving end from multiple different paths. The user is in the area covered by the WiFi signal. During activities, different activities will produce different interferences to multipath signals, which in turn will affect the variation of the channel state information (CSI) amplitude. Therefore, the CSI data is preprocessed and features are extracted, and the correlation between features and action behaviors is used to infer action behaviors by analyzing features to achieve the purpose of recognition. [0003] Smart devices h...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the two problems in the above-mentioned prior art: 1. The low recognition accuracy of the CSI perception system caused by the existence of non-compliant actions; 2. The generalization ability of the recognition model caused by single action distribution and poor data diversity bad question

Images