An efficient built-up land expansion model using a modified U-Net

This paper introduces an improved convolutional neural network based on the conventional U-Net for simulating built-up land expansion. The proposed method hires a pixel-wise semantic segmentation approach considering spatial drivers affecting urbanization as data cubes. Independent variables including altitude, slope, and distance from barren, crop, greenery, roads, urban areas 1998, 2008, 2018 were considered covariates simulation of expansion in Tehran Karaj regions Iran. was compared with random forest (RF) algorithm baseline model. Evaluation using area under total operating characteristic indicated superiority our modified (0.87) over RF (0.82) algorithm. Furthermore, evaluation percent correct metric that model is capable learning neighborhood effects effectively leading to simulate accurately, independent applying cellular automata (CA) Therefore, CA which can consider recommended precisely.