Correcting the geometry and color of digital images

被引：17

作者：

Jackowski, M

Goshtasby, A

Bines, S

Roseman, D

Yu, C

机构：

[1] RUSH PRESBYTERIAN ST LUKES MED CTR,DEPT GEN SURG,CHICAGO,IL 60612

[2] UNIV ILLINOIS,DEPT ELECT ENGN & COMP SCI,CHICAGO,IL 60607

来源：

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE | 1997年 / 19卷 / 10期

基金：

美国国家科学基金会;

关键词：

geometry correction; color correction; approximation; parametric surface; rational Gaussian surface;

D O I：

10.1109/34.625125

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

A unified method for correcting the geometry and color of digital images is presented. The method uses a color chart with a prearranged array of known color patches. Using the true geometry and color of the chart and geometry and color of the image of the chart, transformation functions that describe the geometry and color characteristics of the camera are determined. The transformation functions are then used to correct the geometry and color of obtained images. Experimental results of the proposed method on images acquired under different scene illuminations and camera settings are presented and evaluated.

引用

页码：1152 / 1158

页数：7

共 42 条

[1]

BARNARD K, 1996, P 4 EUR C COMP VIS A, V2, P3

[2]

Cowan W. B., 1983, Computer Graphics, V17, P315, DOI 10.1145/964967.801163

[3]

Duchon J., 1977, CONSTRUCTIVE THEORY, P85, DOI [DOI 10.1007/BFB0086566, 10.1007/BFb0086566]

[4]

ECHIGO T, 1990, MACH VISION APPL, V1, P159