Multi-sensor fusion slam system, multi-sensor fusion method, robot, and medium

A multi-sensor fusion SLAM system operating on a mobile robot, the multi-sensor fusion SLAM system comprising:

• a visual inertia module configured to output pose information;
• a laser scanning matching module configured to use the pose information as an initial value, match a laser scanned point cloud with a voxel map to solve an advanced pose, integrate the point cloud into the voxel map according to the advanced pose, and derive a new voxel subgraph, the laser scanning matching module being configured to a laser matching constraint;
• a loop closure detection module configured to extract an image descriptor of a visual keyframe by using a deep neural network, compare the image descriptor with an image descriptor of a previous keyframe to determine whether a closed loop exists, and if the closed loop exists, determine pose transformation of the two keyframes by using a perspective-n-point (PnP) method, solve a loop closure pose constraint according to the pose transformation and the voxel subgraph, and send the loop closure pose constraint to the laser scanning matching module;
• and
• a visual laser image optimization module configured to correct an accumulated error of the system according to the pose information, the laser matching constraint, and the loop closure pose constraint after the closed loop occurs.