Accelerating a hyperspectral inversion model for submerged marineecosystems using high performance computingon graphical processor units

摘要

Remote sensing of shallow submerged marine ecosystems presents a challenging environment for information extractionalgorithms, where physically based solutions commonly require complex, computationally intensive algorithms. Theinherent variations in water depth, water properties, and surface waves all impact the measured remote sensing signal,and the strong absorption of light in water also limits the effective range of wavelengths available for analysis. Analgorithm has been developed to address this multifaceted problem. The algorithm uses a two-stage inverse semi-analytical optimization model and spectral unmixing scheme to derive water column properties, water depth and habitatcomposition from imaging spectroscopy data. In addition to testing and validation studies, work on this algorithm hasincluded improving its efficiency using the computing power of graphical processor units (GPUs). This improvementprovides accelerated execution of the algorithm, and by leveraging more robust optimization routines, also facilitatesincreased accuracy in algorithm output. Initial results from implementing the algorithm on a single GPU using aconservative optimization strategy indicate substantial improvement in performance can be achieved using thistechnology. We present an overview of the algorithm, provide example output, discuss the GPU parallelizationapproach, and illustrate the performance achievements that have been obtained using GPU technology.