Abstract: To address the issue of low segmentation accuracy in traditional morphological segmenta-tion algorithms under complex map environments, this paper proposed a morphological map segmentation method based on directional erosion and wavefront propagation optimization. The algorithm first introduced a corrosion resistance threshold, which endows the morphological segmentation method with directional erosion. Secondly, by optimizing the number of wavefront propagation iterations based on the erosion count, the undue encroachment of rooms into corridors was significantly mitigated. Furthermore, the algorithm also incorporated a clutter removal step and designed a new heuristic rule to merge over-segmented regions. Experimental evaluation across different environments showed the improved algorithm achieved average F1-scores of 92.2% and 95.3% in scenes with and without furniture, respectively. Furthermore, coverage paths planned using the segmentation results exhibited shorter lengths and fewer turns. These results verify the effectiveness of the proposed approach.