A Seamless Model of the Global Multi-resolution Images Based on QTM

In this paper, a global seamless image model based on Quaternary Triangular Mesh (QTM) is developed. Our approaches start with an Octahedron QTM partition on the spherical surface and an algorithm of transforming square image pixels into QTM pixels is introduced. Next, a Recursive Closing Method is approached to eliminate the cracks caused by different subdivisions. In the end, the experiment is made to test the methods above by using of Wsiearth.tif data with 256 gray levels in ArcGIS. The results illuminate that: (1) this model provides a seamless base for global analysis using global image data and the deficiencies of data broken in traditional model are overcome; (2) the transformation algorithm is feasible and its accuracy can meet most applications requirement; (3) by using the Recursive Closing method, not only the problem of discontinuity between different levels is overcome, but the pixel vertices is ensured to coincide with the vertices of QTM exactly as well.