Small objects detection of hyperspectral image in urban areas

Remote sensing has been widely used for urban planning, urban environment and resources management. Recognition of some small targets is crucial for these purposes, and hyperspectral data can be used to extract targets with its high spectral resolution. This paper presented several representative algorithms of small targets detection in hyperspectral image, and four abnormal detect methods were applied to find vehicles in urban areas. Comparison of presented results is also being conducted to verify that abnormal detecting methods can be used for urban targets suitably and effectively.