Forest canopy structure derived from spatial and spectral high resolution remote sensing data

Abstract

Forest canopy structure can be described by a variety of biophysical parameters, for example leaf area index (LAI) and fractional cover (fcover). These parameters are derived currently from remotely sensed data only with limited accuracy. The retrieval of biophysical parameters is often conducted by empirical models based on vegetation indices (VI) exploiting the spectral information but ignoring the spatial dimension contained in remote sensing data. However, texture information provided by high spatial resolution data can be potentially used as additional information related to the forest structure and might improve the models for the retrieval of biophysical parameters. The aim of this research is to evaluate several methods to combine spectral and textural information to derive the best retrieval method of LAI and fcover from spectral and high spatial resolution remote sensing data in a coniferous forest in the Swiss National Park.
Spectral data as well as spatial data contain information, which can be correlated with the field measurements of biophysical forest parameters. The relationship between spectral data and the field measurements proved to be slightly better than between spatial data and field parameter.