CLASSIFICATION OF SATELLITE IMAGERY WITH TWO APPROACHES OF SPATIAL ATTRACTION (CASE STUDY: PARTS OF HORMOZGAN SHORES)

A wide range of methods have been evaluated for analyzing aerial and satellite images. Remote sensing data are used in many applications. Typically, the process ofimage classification is started to convert data into meaningful information [1]. In this article, two approaches based on the theory of spatial dependence are described.Spatial dependency is known as attracting the near observations more than the distant observations. The first th eory of spatial dependence was presented by Atkinson in1997 [2]. In spatial attraction models, this spatial dependence is interpreted by attraction between neighboring pixels or subpixels [3]. The results show thatthe modified subpixel/spatial attraction model (MSPSAM) with an overall accuracy of 99.83% and kappa of 0.99 has a good performance for classification. The results confirmthe quality of both methods for the classification of satellite imagery.