作者: Juan Pablo Rivera , Jochem Verrelst , Jesús Delegido , Frank Veroustraete , José Moreno
DOI: 10.3390/RS6064927
关键词:
摘要: Regression models based on spectral indices are typically empirical formulae enabling the mapping of biophysical parameters derived from Earth Observation (EO) data. Due to its nature, it remains nevertheless uncertain what extent a selected regression model is most appropriate one, until all band combinations and curve fitting functions assessed. This paper describes application Spectral Index (SI) assessment toolbox in Automated Radiative Transfer Models Operator (ARTMO) package. ARTMO enables semi-automatic retrieval optical remote sensing observations. The SI facilitates parameter accuracy established as well new generic SIs. For instance, formulation used, possible formulations with up ten bands can be defined evaluated. Several options available assessment: calibration/validation data partitioning, addition noise definition models. To illustrate functioning, two-band according simple ratio (SR) normalized difference (ND) various (linear, exponential, power, logarithmic, polynomial) have been HyMap imaging spectrometer (430–2490 nm) obtained during SPARC-2003 campaign Barrax, Spain, used extract leaf area index (LAI) chlorophyll content (LCC) estimates. both SR ND sensitive regions identified for green (539–570 longwave SWIR (2421–2453 LAI (r2: 0.83) far-red (692 NIR (1340 or shortwave (1661–1686 LCC 0.93). Polynomial, logarithmic linear led similar best correlations, though spatial differences emerged when applying imagery. We suggest that systematic strong requirement quality assurance approach accurate retrieval.