Radiometric correction and normalization of airborne LiDAR intensity data for improving land-cover classification

Wai Yeung Yan, Ahmed Shaker

Research output: Journal article publicationJournal articleAcademic researchpeer-review

77 Citations (Scopus)

Abstract

Radiometric correction of airborne LiDAR intensity data has been proposed based on the use of the radar (range) equation for removing the effects of attenuation due to system- and environmental-induced distortions. Although radiometric correction of airborne LiDAR intensity data has been recently investigated with results revealing improved accuracy of surface classification, there exist a few voids requiring further research effort. First, the variation of object surface characteristics (slope and aspect) plays a crucial role in modeling the recorded intensity data, and thus, the laser incidence angle is usually considered in the correction process. Nevertheless, the use of incidence angle would lead to the effects of overcorrection, particularly on those features located in steep slope. Second, line-stripping problems are usually appeared in the overlapping region of LiDAR data strips acquired by sensors configured with automatic gain control (AGC). Currently, the effects of AGC cannot be perfectly modeled due to the nondisclosure of information by the sensor manufacturers. In this paper, we attempt to fill these voids by: 1) proposing a correction mechanism using the surface slope as a threshold to select either using scan angle or incidence angle in the radar (range) equation; and 2) proposing a subhistogram matching technique to radiometrically normalize the overlapping intensity data. The proposed approaches were applied to three real airborne LiDAR data strips for experimental testing. The results showed that the coefficient of variation reached to the lowest value for most of the land-cover features with a slope threshold between 30° and 40°. The variance-to-mean ratio of five land-cover features was significantly reduced by 70%-82% after applying the proposed correction mechanism. In addition, the systematic noises appeared in the overlapping region were significantly reduced after radiometric correction and normalization, where the overall accuracies were improved by up to 16.5% in the results of intensity data classification. With the demonstrated improvement in intensity homogeneity, it is recommended that airborne LiDAR intensity data should be radiometrically preprocessed before performing any thematic applications.

Original languageEnglish
Article number6822559
Pages (from-to)7658-7673
Number of pages16
JournalIEEE Transactions on Geoscience and Remote Sensing
Volume52
Issue number12
DOIs
Publication statusPublished - Dec 2014
Externally publishedYes

Keywords

  • Airborne LiDAR
  • Gaussian mixture model (GMM)
  • incidence angle
  • intensity
  • land-cover classification
  • land-cover homogeneity
  • radiometric correction
  • radiometric normalization
  • subhistogram matching

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • General Earth and Planetary Sciences

Fingerprint

Dive into the research topics of 'Radiometric correction and normalization of airborne LiDAR intensity data for improving land-cover classification'. Together they form a unique fingerprint.

Cite this