A fractal concentration-area method for assigning a color palette for image representation

被引:53
作者
Cheng, QM
Li, QM
机构
[1] York Univ, Dept Earth & Atmospher Sci, Toronto, ON M3J 1P3, Canada
[2] York Univ, Dept Geog, Toronto, ON M3J 1P3, Canada
基金
加拿大自然科学与工程研究理事会;
关键词
fractal; multifractal; color palette; image classification; GIS; ActiveX;
D O I
10.1016/S0098-3004(01)00060-7
中图分类号
TP39 [计算机的应用];
学科分类号
081203 [计算机应用技术]; 0835 [软件工程];
摘要
Displaying the remotely sensed image with a proper color palette is the first task in any kind of image processing and pattern recognition in GIS and image processing environments. The purpose of displaying the image should be not only to provide a visual representation of the variance of the image, although this has been the primary objective of most conventional methods, but also the color palette should reflect real-world features on the ground which must be the primary objective of employing remotely sensed data. Although most conventional methods focus only on the first purpose of image representation, the concentration-area (C-A plot) fractal method proposed in this paper aims to meet both purposes on the basis of pixel values and pixel value frequency distribution as well as spatial and geometrical properties of image patterns. The C-A method can be used to establish power-law relationships between the area A(greater than or equal tos) with the pixel values greater than s and the pixel value s itself after plotting these values on log-log paper. A number of straight-line segments can be manually or automatically fitted to the points on the log-log paper, each representing a power-law relationship between the area A and the cutoff pixel value for s in a particular range. These straight-line segments can yield a group of cutoff values on the basis of which the image can be classified into discrete classes or zones. These zones usually correspond to the real-world features on the ground. A Windows program has been prepared in ActiveX format for implementing the C-A method and integrating it into other GIS and image processing systems. A case study of Landsat TM band 5 has been used to demonstrate the application of the method and the flexibility of the computer program. (C) 2002 Elsevier Science Ltd. All rights reserved.
引用
收藏
页码:567 / 575
页数:9
相关论文
共 27 条
[1]
[Anonymous], 1983, New York
[2]
Geostatistical classification for remote sensing: an introduction [J].
Atkinson, PM ;
Lewis, P .
COMPUTERS & GEOSCIENCES, 2000, 26 (04) :361-371
[3]
Chen CW, 1998, J FOOD DRUG ANAL, V6, P713
[4]
The gliding box method for multifractal modeling [J].
Chen, QM .
COMPUTERS & GEOSCIENCES, 1999, 25 (09) :1073-1079
[5]
Cheng Q., 1999, P 5 ANN C INT ASS MA, V1, P87
[6]
CHENG Q, 1994, THESIS U OTTAWA OTTA
[7]
Cheng Q., 2000, Natural Resources Research, V9, P43, DOI DOI 10.1023/A:1010109829861
[8]
CHENG Q, IN PRESS MATH GEOLOG
[9]
A spatial analysis method for geochemical anomaly separation [J].
Cheng, QM ;
Agterberg, FP ;
BonhamCarter, GF .
JOURNAL OF GEOCHEMICAL EXPLORATION, 1996, 56 (03) :183-195
[10]
Statistical study of trace elements in the soluble organic and amorphous Fe-Mn phases of surficial sediments, Sudbury Basin .1. Multivariate and spatial analysis [J].
Cheng, QM ;
BonhamCarter, GF ;
Hall, GEM ;
Bajc, A .
JOURNAL OF GEOCHEMICAL EXPLORATION, 1997, 59 (01) :27-46