Abstract

The target-visualization method is intended to optimize the process of interpreting hyperspectral images with the user’s interactive participation. This method is a development of the method of visualizing the measure of similarity to a reference and is based on the transformation of a multichannel image into a single-channel one. As a result of the transformation, the value of an attribute (the decision function, defined in terms of the spectral characteristics) is visualized and is used in the interpretation. Such a transformation makes it possible to interactively correct the results of automatic processing simply by changing the brightness range. This paper demonstrates for the first time the combined use of automatic segmentation and target visualization. The possibilities of the method are illustrated, using as an example the solution of problems of contour processing of hyperspectral images and the recognition of objects with indistinctly specified spectral characteristics.

© 2015 Optical Society of America

Multiscale target extraction using a spectral saliency map for a hyperspectral image

Jing Zhang, Wenhao Geng, Li Zhuo, Qi Tian, and Yan Cao
Appl. Opt. 55(28) 8089-8100 (2016)

Feature reduction and morphological processing for hyperspectral image data

David Casasent and Xue-Wen Chen
Appl. Opt. 43(2) 227-236 (2004)

Hyperspectral remote sensing image retrieval system using spectral and texture features

Jing Zhang, Wenhao Geng, Xi Liang, Jiafeng Li, Li Zhuo, and Qianlan Zhou
Appl. Opt. 56(16) 4785-4796 (2017)

Flexible calibration method for visual measurement using an improved target with vanishing constraints

Xiaoyun Chen, Jiarui Lin, Linghui Yang, Yanbiao Sun, and Jigui Zhu
J. Opt. Soc. Am. A 37(3) 435-443 (2020)

Refractive surface flow visualization using image processing

Douglas F. Britton and L. Montgomery Smith
Appl. Opt. 34(11) 1872-1877 (1995)