Human Activity Recognition Method Based on Millimeter Wave Radar Sparse Point Clouds
At present,human behavior recognition methods based on millimeter wave radar cannot distinguish similar actions in complicated scenes.In addition,these methods have the characteristics of low robustness and interference resistance.To address the above two issues,a universal human behavior recognition method based on millimeter wave radar sparse point clouds is proposed.Firstly,the method samples the point cloud using the K-means++clustering algorithm,and then adopts a point cloud activity classi-fication network based on attentional feature fusion for the extraction and recognition of human behavior features,which can consider both the spatial and temporal features of point clouds and has the sensitive perception of sparse point cloud motion.In order to verify the effectiveness and robustness of the proposed method,the experiments are conducted on the MM Activity dataset and MMGesture dataset,respectively,with the accuracy of 97.50％and 94.10％on both datasets,outperforming other methods.Furthermore,the effectiveness of the K-means++point cloud sampling method is further verified,and compared to random sampling,the accuracy is improved by 0.4％.The experimental results show that the proposed method can effectively promote the accuracy of human behavior recognition,and the model has a strong generalization ability.
millimeter-wave radarhuman perceptionbehavior recognitionsparse point cloudsfeature fusion
万方数据
维普
