Seismic Data Representation Based On Subdivided Surfaces

Two typical representations for images are raster graphics and vector graphics.Raster graphics provides an alternative representation for describing images via a grid of pixels, with each pixel storing either a color/grey value or an index into a color palette. The drawbacks of the image representation based on pixel has revealed in many applications such as image magnification and printing. Vector graphics describes images via vector representation which use geometrical primitives such as points, curves and surfaces rather than discrete pixels to represent an image. The vector representation is scalable, compact and resolution independent, which facilitates many image editing processes. Vector-based graphical contents have been widely adopted in personal computers and on the Internet. Raster image vectorization is increasingly significant.The vector representation store small amounts of data since vector graphics is compact. This thesis presents a subdivision-based vector graphics for image representation and creation. The graphics representation is a subdivision surface defined by a triangular mesh with color attribute at patches and feature attribute at edges. The curvilinear features of the image are first detected. And then a meshing algorithm is introduced to divide the segments without edges into sub-regions.Finally, an automatic algorithm is developed to convert a raster image into such a vector graphics representation.In the algorithm edge detector can get an edge first to describe curvilinear features of an image, a Delaunay triangulation is then constructed based on the curvilinear edges, to divide a region into a set of non-overlapping triangles, and finally, Loop subdivision is used to updates the triangulation and the triangular patches color attributes. The automatic vectorization algorithm is applied to the recreation of images based on the seismic data. Numerical results indicate this image vectorization algorithm achieves accurate and compact representation of seismic data with a small amount of coefficients.