Change Detection Using Fast Subpixel Mapping Algorithms for LISS IV Imagery

Abstract

Due to speedy changes on the Earth’s surface, it is vital to perform land cowl amendment detection (CD) at a fine abstraction and fine temporal resolution. However, remote sensing pictures with each fine abstraction and temporal resolutions square measure ordinarily not obtainable or, wherever obtainable, could also be pricey to get. This paper tries to realize fine abstraction and temporal resolution land cowl CD with a brand new technology supported subpixel mapping (SPM). The fine spacial resolution land cowl maps (FRMs) area unit 1st foreseen through SPM of the coarse special, however, fine temporal resolution pictures, and then, subpixel resolution CD is performed by comparison of sophistication labels within the SPM results. For the primary time, 5 quick SPM algorithms, as well as linear interpolation, bicubic interpolation, subpixel/pixel spacial attraction model, Kriging, and radial basis operate interpolation ways, area unit projected for subpixel resolution CD. The auxiliary data from the known FRM on one date is incorporated in SPM of coarse pictures on alternative dates to extend the CD accuracy. Supported the 5 quick SPM algorithms and, therefore, the convenience of the FRM, subpixels for every category area unit expected by comparison of the calculable soft category values at the target fine abstraction resolution and borrowing data from the FRM.