MF-Mamba: Multiscale Convolution and Mamba Fusion Model for Semantic Segmentation of Remote Sensing Imagery

Semantic segmentation of remote sensing imagery plays an important role in applications such as environmental monitoring and disaster response. However, challenges such as complex spatial patterns of variable target objects, significant scale variations, and high interclass similarity challenge accurate segmentation. Most existing methods based on convolutional neural networks (CNNs) and Transformers face limitations in modeling multiscale global–local dependencies or often incur high computational costs. Therefore, we propose a multiscale convolution and Mamba fusion model (MF-Mamba) that integrates a CNN encoder with a Mamba-based decoder. The decoder incorporates a global–local state-space (GLSS) module with eight-directional selective scanning mechanisms and multikernel parallel convolutions to capture the rich global–local context. To enhance multiscale feature representation, we developed a channel–spatial attention and dense multiscale feature fusion (CSDF) module, which combines channel–spatial attention and atrous convolutions for multiscale feature fusion. Additionally, a multiscale lateral connection is developed to align encoder features for efficient integration. Experiments on the datasets of ISPRS Vaihingen, ISPRS Potsdam, and the Wuhan Dense Labeling Dataset (WHDLD) demonstrate the superior performance of MF-Mamba compared to existing state-of-the-art methods. It achieves Mean $F1$ scores of 86.71%, 90.70%, and 77.07%. The code is available at https://github.com/Mango-Mars/MF-Mamba