The Sandia Optical Fringe Analysis Slope Tool (SOFAST) is a tool that has been developed at Sandia to measure the surface slope of concentrating solar power optics. This tool has largely remained of research quality over the past few years. Since SOFAST is important to ongoing tests happening at Sandia as well as an interest to others outside Sandia, there is a desire to bring SOFAST up to professional software standards. The goal of this effort was to make progress in several broad areas including: code quality, sample data collection, and validation and testing. During the course of this effort, much progress was made in these areas. SOFAST is now a much more professional grade tool. There are, however, some areas of improvement that could not be addressed in the timeframe of this work and will be addressed in the continuation of this effort.
In this effort, random noise data augmentation is compared to phenomenologically-inspired data augmentation for a target detection task, evaluated on the Digital Imaging and Remote Sensing Image Generation (DIRSIG) model "MegaScene" simulated hyperspectral dataset. Random data augmentation is commonly used in the machine learning literature to improve model generalization. While random perturbations of an input may work well in certain fields such as image classification, they can be unhelpful in other applications such as hyperspectral target detection. For instance, random noise augmentation may not be beneficial when the applied noise distribution does not match underlying physical signal processes or sensor noise. In the context of a low-noise sensor, augmentation mimicking material mixing and other practical spectral modulations is likely to be more effective when used to train a target detector. It is therefore important to utilize a data augmentation strategy that emulates the natural variability in observed spectra. To validate this claim, a small fully connected neural network architecture is trained using an ideal hemispheric reflectance materials dataset as a trivial baseline. That dataset is then augmented using Gaussian random noise and the model is retrained and again applied to MegaScene. Finally, augmentation is instead performed using phenomenological insight and used to retrain and reevaluate the model. In this work, the phenomenological augmentation implements only simple and commonly encountered spectral permutations, namely linear mixing and shadowing. Comparison is made between the augmented models and the baseline model in terms of low constant false alarm rate (CFAR) performance.