Optimal resolution selection for monitoring forest vegetation using remote sensing images.

Abstract

Image resolution directly affects precision, cost, and efficiency of forest vegetation extraction. To determine the best spatial resolution of remote sensing images for forest vegetation extraction, a method was proposed for optimal spatial resolution of remote sensing images when monitoring forest vegetation. Based on a panchromatic image of GF-2, a step length variation function was used to analyze forest vegetation of three distribution types considering monitoring precision, cost, and efficiency in Yangquan Town, Changning City, Hunan Province. A preliminary minimum resolution of the image suitable for forest vegetation extraction was determined. Then, through resampling of GF-2 multi-spectral images after fusion, middle and low resolution images of different scales were made to extract forest vegetation using the supervised classification method. Quantitative and qualitative analyses of results were carried out for accuracy of forest vegetation extraction, image cost, and data processing time. Results showed that the optimal spatial resolution for remote sensing monitoring images was different for different types of forest vegetation, with forest vegetation of small canopies being 3.2 m, for big canopies being 16.0 m, and for mixed canopies being 8.0 m. The optimal spatial resolution of remote sensing images to monitor forest vegetation could be used as a reference in other areas.