Human Body Movement Posture Monitoring and Recognition Based on KNN Algorithm

Traditional human movement posture monitoring and recognition methods use rule-based methods, but in complex scenes and situations with large posture changes, their accuracy and recall rates pose great challenges. This paper proposes a human movement posture monitoring and recognition method based on the K-Nearest Neighbors (KNN) algorithm. This paper constructed a training data set containing rich posture information, and performed data preprocessing and feature extraction. Then this article calculates the distance between the posture sample to be classified and each sample in the training set, selects K nearest neighbor samples, and performs voting or weighted voting based on the labels of the nearest neighbor samples to determine the posture label of the sample to be classified. Experimental results show that the accuracy of this algorithm remains at a high level of 89%-97%. Compared with traditional methods, this method has achieved significant improvements in accuracy and recall.