U-Net

U-Net is a convolutional neural networkNeural NetworkA Neural Network is a computing system inspired by the structure of biological neural networks in the brain. It forms... architecture originally developed for biomedical image segmentation and since adopted extensively in remote sensingRemote SensingRemote sensing is the science of collecting data about Earth's surface without direct physical contact, primarily usi... and geospatial analysisGeospatial AnalysisGeospatial analysis applies statistical methods and specialized software to interpret spatial data, uncovering patter.... Its distinctive U-shaped architecture consists of a contracting encoder path that captures context through successive downsampling, and an expansive decoder path that enables precise localization through upsampling. Skip connections bridge corresponding encoder and decoder layers, passing high-resolution feature maps directly to the decoder to preserve spatial detail that would otherwise be lost during compression. U-Net for Geospatial Segmentation TasksU-Net has become the default architecture for many pixel-level geospatial prediction tasks. Building footprintBuilding FootprintA Building Footprint is the outline of a building as seen from directly above, representing the area of ground it occ... extraction uses U-Net to delineate individual structures from high-resolution aerial and satellite imagerySatellite ImagerySatellite imagery consists of photographs and data captured by Earth observation satellites orbiting the planet. Thes.... Road network extraction traces road surfaces at pixel resolution for map generation. Flood extent mapping applies U-Net to rapidly delineate inundated areas from optical or SARSARSynthetic Aperture Radar (SAR) is an active remote sensing technology that uses microwave radar pulses to create high... imagery during disaster response. Agricultural field boundary detection segments individual parcels for precision farming applications. Cloud and shadow masking uses U-Net to identify contaminated pixels in optical satellite imagery for data quality assurance. Variants and Practical AdvantagesU-Net's architecture works well with limited training data, which is a critical advantage in geospatial applications where labeled datasets are expensive to create. The skip connections ensure that fine boundary details are preserved, producing cleaner segmentation maps than architectures without them. Popular variants include Attention U-Net, which adds attention gates to focus on relevant features, and U-Net++, which uses nested dense skip connections for improved feature fusion. Residual U-Net incorporates residual blocks for deeper networks that train more stably. These variants consistently achieve top performance in geospatial segmentation benchmarks.