CMLFormer: CNN and Multiscale Local-Context Transformer Network for Remote Sensing Images Semantic Segmentation

The characteristics of remote sensing images, such as complex ground objects, rich feature details, large intraclass variance and small interclass variance, usually require deep learning semantic segmentation methods to have strong feature learning representation ability. Due to the limitation of convolutional operation, convolutional neural networks (CNNs) are good at capturing local details, but perform poorly at modeling long-range dependencies. Transformers rely on multihead self-attention mechanisms to extract global contextual information, but it usually leads to high complexity. Therefore, this article proposes CNN and multiscale local-context transformer network (CMLFormer), a novel encoder-decoder structured network for remote sensing image semantic segmentation. Specifically, for the features extracted by the lightweight ResNet18 encoder, we design a transformer decoder based on multiscale local-context transform block (MLTB) to enhance the ability of feature learning. By using a self-attention mechanism with nonoverlapping windows and with the help of multiscale horizontal and vertical interactive stripe convolution, MLTB is able to capture both local feature information and global feature information at different scales with low complexity. In addition, the feature enhanced module is introduced into the decoder to further facilitate the learning of global and local information. Experimental results show that our proposed CMLFormer exhibits excellent performance on the Vaihingen and Potsdam datasets.