Remote sensing image retrieval based on multi-channel Gabor texture features.
Lu Li, Gis K · Journal of Zhejiang University(Sciences Edition) · 2004
The rapid increase in remote sensing image data has posed a tough task of retrieving and managing image efficiently for researchers. A novel remote sensing image retrieval approach based on multi-channel Gabor texture features and quin-tree decomposition is proposed. A given image is decomposed by using quin-tree, and texture features of each subimage are extracted via multi-channel Gabor filter. The similarity between a given query image and the database subimages, whose sizes are approximately equal to the former, is measured by using Euclid texture distance. Then the most similar subimages are returned as query results. This approach is applied to retrieve high resolution remote sensing images, and its efficiency is confirmed by the experiments.