基于DDPM-MBN的井下人员步态识别方法

打开文本图片集

中图分类号：TD67 文献标志码：A

Abstract：Existing milimeter-wave radar-based gait recognition methods are typically trained on small-scale datasets，resulting in poor model generalization and limited ability to extract efective globaland local features from complex underground environments，which leads to low recognition accuracy.Toaddress these issues，agait recognition method for underground personnel based on the Denoising Diffusion Probabilistic Model（DDPM） and Multi-Branch Network （MBN） was proposed.The DDPM was used to denoise and augment the time-frequencyspectrograms converted fromradar echoes，effectively expanding the quantityofunderground gait dataandimproving dataquality.TheMBN，consistingofoneglobal branchand two localbranches，extracted global gait features and local features of diferent granularities，enabling sufficient multi-scale feature extraction and improving the recognition of walking directionand speed.The Softmax loss and triplet loss were jointly employed tooptimize coarse-grained features （2 O48-dimensional features before dimensionalityreduction）and fine-grained features （256-dimensional features afterdimensionality reduction）inacollborative manner，thereby enhancing the model'soverallclassificationabilityand feature discriminability.Experimentalresultsshowed that， on the self-built gait dataset，the DDPM-MBN model achieved Rank-1 accuracyand mean Average Precision （mAP） improvements of 8.05% and 16.96% ，respectively，compared with ResNet-50.Compared with mainstream gait recognition models，the DDPM-MBN model achieved the best performance，with Rank-1 accuracyand mAP reaching 97.91% and 95.48% ，respectively.

Key words：gaitrecognition;millimeter-waveradar;denoising difusion probabilistic model;multi-branch network; time-frequency spectrogram

0引言

在智能化矿山建设中，实现井下作业人员身份识别是提升安全管理效率的关键[1]。（剩余10181字）