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.spidertattooz.com/iOptron-HAE43EC-Strain-Wave-AZ-EQ-Mount-w-iPolar-High-Precision-Encoder/