A new sub-pixel mapping method based on spatial autocorrelation and landscape indexes

Geospatial Technology is one of the three emerging technologies in the 21st century. Remote sensing images typically contain a combination of pure and mixed pixels. Mixed pixels result when the sensor's instantaneous field-of-view (IFOV) includes more than one land cover class on the ground. The phenomenon of this "mixed pixels" caused great difficulties to remote sensing image classification. Moreover, it serious impacts on the accuracy and effectiveness of the results of remote sensing image classification. It has become to the major issue which to obstruct the quantitative of remote sensing technology in-depth development. These mixed pixels pose a difficult problem for RS classification, as their spectral characteristics are not representative of any single land cover class. In fact, the value of each pixel is the composite spectral signature of the land cover types present. For these mixed pixels, fuzzy classifiers can be used, which assign a pixel to several land cover classes in proportion to the area of the pixel that each class covers. Fuzzy techniques aim to estimate the proportions of specific classes that occur within each pixel. The result is a number of fraction images, one for each land cover class concerned. While this information describes the class composition, it does not provide any indication as to how the classes are spatially distributed within the pixel. A limited number of methods for solving this sub-pixel mapping problem have been proposed. Schneider (1993) introduced a knowledge-based analysis technique for automatic localization of field boundaries in scenes of agricultural areas. It is applicable to homogeneous fields with straight boundaries. Gavin and Jennison (1997) adopted a Bayesian approach and incorporated prior information about the true image in a stochastic model that attaches higher probability to images with shorter total edge length. Atkinson (1997)