Author_Institution :
Sch. of Biol. Sci. & Technol., Hubei Inst. for Nat., Enshi, China
Abstract :
The QuickBird phantom was the multi-spectrum resolution has achieved Mi Ji one of commercial satellites, might supply the spectrum information, the structure information and the time information for the user, but in the actual application by the spectrum, the time information were many, the structure information were few, but the structure information´s application increased the remote sensing phantom classified precision. The structure information´s extraction uses in statistics variation function method, when the computation variation function uses the overlaid windows, the size is the 5×5 size, the length of stride is 1. But the QuickBird phantom´s multi-spectra have 4 wave bands, to reduce the data redundancy, before its extraction texture uses the main ingredient transformation, the first host ingredient information content is 67.04%, the second host ingredient is 32.45%, one, two main ingredient sum total is 99.49%, according to the above only withdraws the main ingredient 1, 2 image texture. Is best after the quite absolute deviation function texture.
Keywords :
feature extraction; image resolution; image texture; remote sensing; statistical analysis; QuickBird image; computation variation function; data redundancy; geostatistics; image texture; multispectrum resolution; remote sensing phantom; spectrum information; statistics variation function method; structure information; texture extraction; time information; Biology; Data mining; Digital images; Image analysis; Image texture analysis; Imaging phantoms; Information analysis; Pattern analysis; Remote sensing; Satellites; QuickBird phantom; Texture; geostatistics;