|
چکیده
|
Forest ecosystems play an important role in the global carbon cycle and it is largely unknown how this role might be altered by transients imposed by global change and deforestation. Remote sensing can provide information on ecosystem state and functioning and, among others, two remote sensing techniques, airborne laser scanning (ALS) and imaging spectroscopy (IS), have been used to characterize forest ecosystems, both independently and combined in fusion approaches. However, the fusion of these datasets should make the best use of the complementarity of both sensors and provide better and more robust vegetation products in forested ecosystems. Similar to other data fusion approaches, satisfying results depend on choosing appropriate fusion levels and methods. In this review paper, we summarize and classify relevant studies that focused on forest characterization using combined ALS and IS data, limited to the last decade. We classified the approaches by fusion level (data or product level) and by choice of methods (physical or empirical methods). Five different categories of products (landcover maps, aboveground biomass, biophysical parameters, gross/net primary productivity and biochemical parameters), have been found as the main aspects of forest ecosystems studied so far. A qualitative accuracy analysis of the products exposed that currently landcover maps are profiting the most from ALS and IS data fusion, while there is room for improvements in respect to the other products, such as biophysical parameters. Only few studies using physical approaches were found, but we expect the use of such approaches will increase with the growing availability of physically based radiative transfer models that can simulate both, ALS and IS data.
|