Abstract
Accurate and efficient simulation of remote sensing images is increasingly needed in order to better exploit remote sensing observations and to better design remote sensing missions. DART (Discrete Anisotropic Radiative Transfer), developed since 1992 based on the discrete ordinates method (i.e., standard mode DART-FT), is one of the most accurate and comprehensive 3D radiative transfer models to simulate the radiative budget and remote sensing observations of urban and natural landscapes. Recently, a new method, called DART-Lux, was integrated into DART model to address the requirements of massive remote sensing data simulation for large-scale and complex landscapes. It is developed based on efficient Monte Carlo light transport algorithms (i.e., bidirectional path tracing) and on DART model framework. DART-Lux can accurately and rapidly simulate the bidirectional reflectance factor (BRF) and spectral images of arbitrary landscapes. This paper presents its theory, implementation, and evaluation. Its accuracy, efficiency and advantages are also discussed. The comparison with standard DART-FT in a variety of scenarios shows that DART-Lux is consistent with DART-FT (relative differences <1%) with simulation time and memory reduced by a hundredfold. DART-Lux is already part of the DART version freely available for scientists ( https://dart.omp.eu ). • DART-Lux: new DART Monte Carlo method for accurate and fast simulation of RS images. • DART-Lux theory, implementation and evaluation. • Theoretical analysis of variance (Monte Carlo noise) in DART-Lux images. • Sensitivity study of performance (accuracy, efficiency) of DART-Lux.
Accepted Version (
Free)
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call