Target coherent action detection method and system based on millimeter wave radar point cloud network and long and short time model

The invention discloses a target coherent action detection method and system based on a millimeter wave radar point cloud network and a long-short time model, and the method comprises the steps: firstly obtaining radar point cloud data collected from a detection region, and determining that the data are arranged according to a time sequence; the actions are classified and stored; inputting a certain type of action, judging the employment of the frame according to the number of point clouds, and inputting the frame to a point cloud network for sampling and grouping; a 100 * 6 feature matrix is output after passing through a full connection layer, and enters a loss function of the feature matrix, so that a more accurate feature matrix can be obtained later; storing the feature matrix with the size of 100 * 6 into a time sequence matrix with the size of 100 * 10 * 6 according to a certain sequence; inputting every ten time sequence matrixes as a Batch into the long-short time model, and entering a loss function