SLAM mapping method based on water surface multi-sensor fusion

The invention discloses a water surface multi-sensor fusion SLAM mapping method applied to water surface unmanned driving. The method comprises the following steps: acquiring positioning information GPS of a robot through a GNSS sensor; acquiring acceleration information IMU of the robot in each direction through an IMU combined positioning module; obtaining a visual odometer VO through a binocular camera; the positioning information GPS, the acceleration information IMU and the visual odometer VO are fused through the constructed linear filtering model of the integrated navigation system, and fused pose information is obtained; positioning the position of the robot according to the distance information and the fused pose information, and constructing an initial map image; and optimizing the initial map image through a visual SLAM technology to generate a two-dimensional grid map. According to the method, a multi-sensor data fusion method is utilized to overcome the difficulty that a single sensor cannot real