Existing road extraction methods based on deep learning often struggle with distinguishing ground objects that share similar spectral information. such as roads and buildings. Consequently. this study proposes a dual encoder-decoder deep neural network to address road extraction in complex backgrounds. https://www.coraldinernyc.com/product-category/hoodie/