Remote-Sensing Image Classification

Remote-sensing image classification involves automatically identifying and labeling what appears in satellite or aerial imagery — forests, urban areas, bodies of water, agricultural land — by analyzing the light reflected or emitted from Earth's surface across many wavelengths. Hyperspectral sensors push this further by capturing hundreds of narrow spectral bands, enabling finer material distinctions than the human eye or conventional cameras can resolve, though the resulting data volumes and complexity demand sophisticated processing. Deep learning has become central to the effort, improving how models extract meaningful spatial and spectral features from these images, while tasks like change detection — spotting how a landscape has shifted between two time points — and spectral unmixing — separating mixed signals from pixels that capture multiple land types at once — remain active and demanding problems. A persistent open question is how to build models that generalize reliably across different sensors, geographies, and acquisition conditions without requiring vast amounts of labeled training data from each new setting.

Works

65,639

Total citations

1,239,336

Keywords

HyperspectralImage AnalysisRemote SensingClassificationDeep LearningChange Detection