Skip to main navigation Skip to search Skip to main content

Reflectance and Surface Normals from a Single Hyperspectral Image

  • Wenhao Xiang
  • , Xudong Jin
  • , Yanfeng Gu*
  • *Corresponding author for this work
  • China State Shipbuilding Corporation
  • Harbin Institute of Technology

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Hyperspectral imaging, which collects rich spectral and spatial information, is a powerful Earth vision method and has many applications. As the data structure is highly complex, the key problem of hyperspectral image processing is in extracting the useful information we want. Traditional feature extraction methods are designed to this end; however, they undergo severe limitations. Most of them are designed mathematically instead of physically and ignore the fact that the changes in physical imaging conditions have a significant influence on the spectra intensity observed. In this chapter, we try to analyze the information contained in hyperspectral images (HSIs) from the perspective of the hyperspectral imaging principle, and propose a novel method of extracting reflectance and surface normals from HSIs.

Original languageEnglish
Title of host publicationThe Proceedings of the International Conference on Sensing and Imaging, 2018
EditorsEric Todd Quinto, Nathan Ida, Ming Jiang, Alfred K. Louis
PublisherSpringer
Pages105-113
Number of pages9
ISBN (Print)9783030308247
DOIs
StatePublished - 2019
Event4th International Conference on Sensing and Imaging, ICSI 2018 - Liuzhou, China
Duration: 15 Oct 201818 Oct 2018

Publication series

NameLecture Notes in Electrical Engineering
Volume606
ISSN (Print)1876-1100
ISSN (Electronic)1876-1119

Conference

Conference4th International Conference on Sensing and Imaging, ICSI 2018
Country/TerritoryChina
CityLiuzhou
Period15/10/1818/10/18

Keywords

  • Hyperspectral
  • Reflectance
  • Surface Normals

Fingerprint

Dive into the research topics of 'Reflectance and Surface Normals from a Single Hyperspectral Image'. Together they form a unique fingerprint.

Cite this