Mapping pure mangrove patches in small corridors and sandbanks using airborne hyperspectral imagery

Cheng Chien Liu, Tsai Wen Hsu, Hui Lin Wen, Kung Hwa Wang

Research output: Contribution to journalArticlepeer-review

4 Citations (Scopus)

Abstract

Taijiang National Park (TNP) of Taiwan is the northernmost geographical position of mangrove habitats in the Northern Hemisphere. Instead of occupying a vast region with a single species, the mangroves in TNP are usually mingled with other plants in a narrow corridor along the water or in groups on a small sandbank. The multi-spectral images acquired from the spaceborne platforms are therefore limited in mapping the abundance and distribution of the mangrove species in TNP.We report the work of mapping pure mangrove patches in small corridors and sandbanks in TNP using airborne Compact Airborne Spectrographic Imager (CASI) hyperspectral imagery. Bu considering the similarity of spectral reflectance among three species of mangrove and other plants, we followed the concept of supervised classification to select a few training areas with known mangrove trees, where the training areas are determined from the detailed map of mangrove distribution derived from the field investigation. The Hourglass hyperspectral analysis technique was employed to identify the endmembers of pure mangrove in the training areas. The results are consistent with the current distribution of mangrove trees, and the remarkable feature of a "mangrove desert" highlights a fact that biodiversity can be easily and quickly destroyed if no protection is provided. Some remnant patches located by this research are very important to the management of mangrove trees.

Original languageEnglish
Article number592
JournalRemote Sensing
Volume11
Issue number5
DOIs
Publication statusPublished - 2019 Mar 1

All Science Journal Classification (ASJC) codes

  • Earth and Planetary Sciences(all)

Fingerprint

Dive into the research topics of 'Mapping pure mangrove patches in small corridors and sandbanks using airborne hyperspectral imagery'. Together they form a unique fingerprint.

Cite this