| Autor: |
Rasel, Sikdar M. M., Hsing-Chung Chang, Diti, Israt Jahan, Ralph, Tim, Saintilan, Neil |
| Předmět: |
|
| Zdroj: |
ISPRS Annals of Photogrammetry, Remote Sensing & Spatial Information Sciences; 2016, Vol. 3 Issue 8, p85-92, 8p |
| Abstrakt: |
Saltmarsh is one of the important communities of wetlands. Due to a range of pressures, it has been declared as an EEC (Ecological Endangered Community) in Australia. In order to correctly identify different saltmarsh species, development of distinct spectral characteristics is essential to monitor this EEC. This research was conducted to classify saltmarsh species based on spectral characteristics in the VNIR wavelength of Hyperion Hyperspectral and Worldview 2 multispectral remote sensing data. Signal Noise Ratio (SNR) and Principal Component Analysis (PCA) were applied in Hyperion data to test data quality and to reduce data dimensionality respectively. FLAASH atmospheric correction was done to get surface reflectance data. Based on spectral and spatial information a supervised classification followed by Mapping Accuracy (%) was used to assess the classification result. SNR of Hyperion data was varied according to season and wavelength and it was higher for all land cover in VNIR wavelength. There was a significant difference between radiance and reflectance spectra. It was found that atmospheric correction improves the spectral information. Based on the PCA of 56 VNIR band of Hyperion, it was possible to segregate 16 bands that contain 99.83% variability. Based on reference 16 bands were compared with 8 bands of Worldview 2 for classification accuracy. Overall Accuracy (OA)% for Worldview 2 was increased from 72 to 79 while for Hyperion, it was increased from 70.47 to 71.66 when bands were added orderly. Considering the significance test with z values and kappa statistics at 95% confidence level, Worldview 2 classification accuracy was higher than Hyperion data. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
