This thesis focuses on the analysis of 3D digital images through the topology theory on the digital space Z3.First of all,a topological structure on Z3 is established.It is called GK-topology which is the product topology of the grid point topology(briefly,GP2-topology)on the 2D digital space Z2 and the Khalimsky line topology(briefly,K1-topology).The smallest open neighborhood structure of every point related to GK-topology is analyzed.The points in digital space Z3 are classified according to the smallest open neighborhood structures.Then some drawbacks of GK-continuous map and GK-homeomorphism are dis-cussed in rotating and classifying digital images.In order to overcome these drawback-s,the thesis introduces the notions of GK-adjacency neighborhood and GK-adjacency set,and so establishes a new relation among pixels.Furthermore,two new maps,GK-adjacency map and GK-A-map,are defined and it is proved that the GK-A-map is a connectedness-preserving map.The difference between a GK-A-map and a GK-continuous map is made clear by comparison and so do GK-adjacency map and GK-continuous map.In fact,a GK-continuous map must be a GK-A-map but the converse is not all true.Based on the GK-A-map and GK-continuous map,two new categories,GK-adjacency category(GKAC)and GK-topology category(GKTC),are established.The thesis al-so introduces a GK-A-isomorphism,and proves that a GK-homeomorphism must be a GK-A-isomorphism,but not vice versa.A classification of 3D digital image is realized by means of a GK-A-isomorphism.Finally,the thesis provides a method of thinning or compressing digital images through GK-A-retraction,which is based on the GK-topology.The method could be used in computer image analysis and image processing. |