Moving object real-time detection algorithm based on three-dimensional laser point cloud

The invention discloses a moving object real-time detection algorithm based on a three-dimensional laser point cloud. The moving object real-time detection algorithm comprises the steps of: (1) outputting position information of points scanned in an FOV space by means of a three-dimensional laser sensor, and converting all the output data into a standard point cloud data format; (2) recording a plurality of frames of original point cloud data, and accumulating the plurality of frames of original point cloud data to obtain a piece of background point cloud data; (3) acquiring real-time point cloud data according to the original point cloud data; (4) comparing the real-time point cloud data with the background point cloud data to obtain differential point cloud data; (5) and assembling the differential point cloud data, and carrying out clustering processing according to distances among points separately to obtain a plurality of sets of clustering points, wherein each set represents a moving object. The moving