Fig. 1. Overall framework of EF-SLAM method

Fig. 2. Comparison of the extraction results using the raster map method and the actual situation. (a) Waterline points extracted by the raster map method; (b) actual waterline points

Fig. 3. Front-view projection point cloud elevation visualization

Fig. 4. Schematic of the criteria for identifying waterside feature points

Fig. 5. Schematic of feature point association method. (a) Waterside feature points; (b) plane feature points

Fig. 6. Factor graph of EF-SLAM method

Fig. 7. Unmanned boat system platform

Fig. 8. EF-SLAM mapping results and comparison of motion trajectories. (a) EF-SLAM mapping result for N03_3; (b) EF-SLAM mapping result for N03_2; (c) comparison of motion trajectories for N03_3; (d) comparison of motion trajectories for N03_2

Fig. 9. Global distribution of absolute trajectory errors. (a) A-LOAM; (b) LEGO-LOAM; (c) LIO-SAM; (d) EF-SLAM

Fig. 10. Global distribution of relative trajectory errors. (a) A-LOAM; (b) LEGO-LOAM; (c) LIO-SAM; (d) EF-SLAM

Fig. 11. Experimental scene diagram

Fig. 12. N02_4 side view of mapping results. (a) A-LOAM; (b) LEGO-LOAM; (c) LIO-SAM; (d) EF-SLAM

Fig. 13. Satellite map of surveyed dataset

Fig. 14. Overhead view of mapping results of surveyed dataset. (a) A-LOAM; (b) LEGO-LOAM; (c) LIO-SAM; (d) EF-SLAM

Fig. 15. Side view of mapping results of surveyed dataset. (a) A-LOAM; (b) LEGO-LOAM; (c) LIO-SAM; (d) EF-SLAM