Position recognition method and device based on sparse point group using low channel 3D lidar sensor

摘要

The present invention relates to a sparse point group-based location recognition method and apparatus using a low-channel three-dimensional lidar sensor capable of increasing accuracy at low cost. The present invention includes: a point cloud collecting unit for collecting point cloud data according to the position of an object detected by a lidar sensor; a point cloud matching unit for detecting the movement of the autonomous vehicle by matching the point cloud data for the past time and the point cloud data for the current time among the collected point cloud data; a point cloud accumulator for repeatedly accumulating the point cloud data for each preset unit distance based on the detected movement of the vehicle for a preset reference number of times; a point cloud filtering unit configured to filter the accumulated point cloud data based on a voxel grid so that a single point exists in a voxel having a preset size; and a location recognition unit that acquires location information of the autonomous vehicle from GPS and recognizes the location of the autonomous vehicle on the global map by matching the filtered point cloud data with the global map based on the location information. Collecting point cloud data according to the location of an object through a sparse point cloud-based position recognition device using a low-channel 3D lidar sensor and a lidar sensor; detecting the movement of the autonomous vehicle by matching the point cloud data for the past time and the point cloud data for the current time among the collected point cloud data; repeatedly accumulating the point cloud data for each predetermined unit distance based on the detected movement of the autonomous vehicle for a predetermined reference number of times; filtering the accumulated point cloud data based on a voxel grid so that a single point exists in a voxel having a preset size; and recognizing the location of the autonomous vehicle by matching the filtered point cloud data with a global map based on the location information of the autonomous vehicle obtained from GPS. A sparse point cloud-based location recognition method is the technical gist.