Vegetation index analysis of multi-source remote sensing data in coal mine wasteland

Abstract

Thirty-six soil samples were collected and their hyperspectral data used to calculate vegetation indices such as a normalised difference vegetation index (NDVT) and a difference vegetation index (DVT). These were evaluated for typical surface object features within the wastelands around Haizhou Opencast Coal Mine in Fuxin city. Aprincipal component analysis to the hyperspectral data was performed, and the result showed that the first and the second principal components satisfactorily accounted for the multi-spectral image information. The panchromatic and multi-spectral images of SPOT5 were then merged. The panchromatic image replaced the first principal component to improve spatial resolution of the image. In addition, the multispectral images and the NDVT image were classified into six types using the unsupervised classification method. The linear quantitative models were built up and the highest correlation coefficients were obtained between the hyperspectral vegetation index and the vegetation index data from the SPOT5 image. The results show that the hyperspectral data and remote sensing images can be used for quantitative estimation of soil nutrients in coal mine wasteland. They can also provide large area surface information for fast and effective decision making regarding revegetation and the monitoring of dynamic change. © 2007 Taylor & Francis Group, LLC.