Open AccessAn Evaluation of Several Pansharpening Methods for Mapping Quasi-circular Vegetation Patches Using GF-2 Imagery
Author(s) -
Qingsheng Liu
Publication year - 2020
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/790/1/012104
Subject(s) - hue , remote sensing , principal component analysis , sharpening , vegetation (pathology) , pattern recognition (psychology) , artificial intelligence , computer science , environmental science , geography , medicine , pathology
It is important to identify the quasi-circular vegetation patches (QVPs), which will help understanding the local ecosystem structure, function, evolution, and maintenance, and is beneficial to make a vegetation restoration in the Yellow River Delta, China. The fused imagery with high spectral and spatial resolution are most appropriate data for mapping the QVPs. This study compared the widely used pansharpening approaches such as the modified intensity-hue-saturation, Gram-Schmidt, colour spectral sharpening, and principal component analysis (PC) approach for mapping the QVPs using the tasselled cap brightness and greenness components of one scene of the spring GF-2 imagery with the decision tree classifier. Overall, the PC method produced a slightly good detection result of the QVPs over the other three pansharpening methods. However, the detection accuracy was still low (F measure = 56.8%). It could be improved using multitemporal images and patch splitting techniques in the future.