Contents Abstract ----------------------------------------------------------------- i Acknowledgments ---------------------------------------------------------- v Contents ---------------------------------------------------------------- vi List of Figures ---------------------------------------------------------- x List of Tables --------------------------------------------------------- xii List of Symbols --------------------------------------------------------- xv 1 Introduction -------------------------------------------------------------- 1 1.1 Purpose of this research -------------------------------------------- 1 1.2 Background of this research ----------------------------------------- 3 1.2.1 History of Hilbert curve -------------------------------------- 3 1.2.2 Applications of the HH-expression to image processing and data compression for binary data ------------------------------ 5 1.2.3 Application to multi-valued images coding using bit-plane decomposition ------------------------------------------------- 6 1.2.4 Application to classifying LANDSAT multi-spectral image data -- 8 1.3 Contents of the thesis ---------------------------------------------- 9 2 Expression of N-dimensional data using a Hilbert scan -------------------- 11 2.1 Definition of the HH-expression ------------------------------------ 11 2.2 Implementation of a Hilbert address ordering algorithm in R2-------- 17 2.2.1 Address alignment and a recursive algorithm in R2------------- 17 2.2.2 Fast algorithm for Hilbert address ordering in R2------------- 20 2.2.3 Hardware implementation for Hilbert address ordering in R2---- 22 2.3 Implementation of a Hilbert address ordering algorithm in R3-------- 25 2.3.1 Recursive algorithm in R3------------------------------------- 25 2.3.2 Fast algorithm for Hilbert address ordering in R3------------- 28 2.3.3 Hardware implementation for Hilbert address ordering in R3---- 31 2.4 Implementation of a Hilbert address ordering algorithm in RN-------- 31 2.4.1 Hilbert address order in RN----------------------------------- 31 2.4.2 Description of Hilbert address orders in RN------------------- 32 2.4.3 Method for constructing look-up tables of terminal rules ----- 33 2.4.4 Look-up tables of induction rules ---------------------------- 37 2.4.5 Fast algorithm for Hilbert address ordering in RN------------- 44 2.4.6 Discussions -------------------------------------------------- 47 3 Applications to image processing and data compression for binary images -- 50 3.1 HH-expressions in image processing --------------------------------- 50 3.1.1 DF-expressions ----------------------------------------------- 50 3.1.2 Difference between the Hilbert scan and the Morton scan on the quadtree ----------------------------------------------------- 51 3.1.3 HH-expression for binary images ------------------------------ 51 3.1.4 Displaying a picture from the HH-expression ------------------ 54 3.1.5 Traversal algorithm for the HH-expression -------------------- 55 3.1.6 Previewing images -------------------------------------------- 57 3.1.7 Circular shifting -------------------------------------------- 58 3.1.8 Rotation by multiples of 90°and mirror operations ----------- 60 3.1.9 Magnification and reduction ---------------------------------- 61 3.1.10 Spectrum of a picture ---------------------------------------- 61 3.1.11 Logical operations on the HH-expression ---------------------- 61 3.2 Compression technique for long term storage ------------------------ 62 3.2.1 Run-length coding from the HH-expression --------------------- 63 3.2.2 Experiments for two-dimensional data ------------------------- 64 3.2.3 Experiments for three-dimensional data ----------------------- 68 4 Application to the compression of gray image data ------------------------ 73 4.1 Problem of the bit-plane decomposition for data compression -------- 73 4.2 Property of bit-plane decomposition -------------------------------- 74 4.2.1 Property of bit-planes and natural digital images ------------ 74 4.2.2 Some properties of bit-planes and real digital images -------- 77 4.2.3 Where and how should we uniformalize? ------------------------ 81 4.3 Uniformalization process using the HH-expression on each bit-plane - 82 4.3.1 Complexity measures on the HH-expression --------------------- 83 4.3.2 Measure for the uniformalization process using the HH-expression -------------------------------------------- 83 4.3.3 Hierarchical uniformalization process ------------------------ 85 4.4 Algorithm for the hierarchical uniformalization process ------------ 87 4.4.1 Variables and a stack operation ------------------------------ 88 4.4.2 Components of the uniformalization process ------------------- 91 4.5 Experimental results ---------------------------------------------- 100 4.5.1 Simulation of other methods --------------------------------- 101 4.5.2 Results and discussions ------------------------------------- 102 5 Application to LANDSAT image classification ----------------------------- 107 5.1 N-dimensional image expressions ----------------------------------- 107 5.2 Hierarchical histogram expression using a Hilbert scan ------------ 110 5.2.1 Hierarchical histogram expression --------------------------- 110 5.2.2 Problem of the division ------------------------------------- 113 5.3 Data structure of the tree structure expression ------------------- 114 5.3.1 List-like data structure ------------------------------------ 115 5.3.2 Image processing on this data structure --------------------- 116 5.3.3 Data compression of this data structure --------------------- 117 5.4 Interactive analysis for multi-spectral image classification ------ 117 5.4.1 LANDSAT-5 TM image data ------------------------------------- 117 5.4.2 Classification problem for multi-spectral images ------------ 118 5.4.3 Overview of our method -------------------------------------- 119 5.4.4 Constructing data tables ------------------------------------ 119 5.4.5 Interactive system for LANDSAT-5 TM image classification ---- 121 5.5 Experiments ------------------------------------------------------- 125 5.5.1 Experimental system ----------------------------------------- 125 5.5.2 System evaluations ------------------------------------------ 125 5.5.3 Discussions ------------------------------------------------- 126 6 Conclusions ------------------------------------------------------------- 129 6.1 HH-expressions ---------------------------------------------------- 129 6.2 Implementation of the Hilbert address ordering algorithm ---------- 129 6.3 Application to image processing and data compression for binary images ------------------------------------------------- 130 6.4 Application to data compression of gray images -------------------- 130 6.5 Application to LANDSAT image classification ----------------------- 131