A Computation Reduced Technique to Primitive Feature Extraction for Image Information Mining Via the Use of Wavelets

NASA’s Earth Station receives terabytes of data daily from satellites. However, the usability of the vast amount of data is constrained to users of different domains because data accessibility is limited to queries based on geographical coordinates, time of acquisitions, sensor types, and acquisition mode. Research in knowledge discovery from remote sensing archives has been propelled in the last few years by the adoption of newer technologies. However, processes for archiving, information retrieval, and distribution of different kinds of imagery received from different satellites are still in their infancy. In general, the imagery received from each satellite provides different information that depends on spectral resolution, spatial resolution, temporal resolution, number of bands, and other characteristics that pose greater challenges for image information mining in remote sensing archives then in multimedia image archives. Hence, the adoption of new technologies that allow the accessibility of remote sensing data based on content and semantics is required to overcome this challenge and increase useful exploitation of the data [1, 2, 3].