Translator Disclaimer
8 September 2006 Spectral image improvement analysis of the model-based spectral imaging reconstruction algorithm
Author Affiliations +
Previous papers introduced a method for simultaneously improving the spatial and spectral resolution of spectral images by developing a model of the spectral sensor and using this model in a statistical minimization algorithm. This paper expands on the Model-Based Spectral Image Reconstruction (MBSIR) algorithm by analyzing the lower bounds on the algorithm's performance. While the MBSIR algorithm improves both spatial and spectral resolution, just the spectral bounds will be analyzed in this paper. This is a valid approach since the functions describing the spatial and spectral blurring are separable. Two lower bounds will be analyzed. The first is the lower bound on spectral resolution and the second is on the spectral accuracy. The spectral resolution lower bound analyzes the improvement the MBSIR algorithm can achieve in resolving two closely spaced spectral features. The spectral accuracy lower bound analyzes the ability of the MBSIR algorithm to reconstruct a spectral feature at the correct location. The sensor model used for this analysis is the AEOS Spectral Imaging Sensor (ASIS). ASIS is located at the Maui Space Surveillance Complex (MSSC) and is used to collect spectral images of space objects. Since all of the objects that ASIS images are non-stationary, the bounds can be used to determine a filter sampling that balances imaging time and image enhancement through the development of a Spectral Reconstruction Capability Metric (SRCM). The SRCM is important for the operation of ASIS. ASIS collects one spectral image at a time to create the spectral image cube. Since ASIS is intended to image space objects that are in orbit, delays in collecting the entire spectral image cube could result in an orientation change in the object. The orientation change could prevent the MBSIR algorithm from working on the data. The SRCM provides a method for determining the optical collection parameters to minimize object motion while maintain algorithm performance. The SRCM also allows for a way to compare different parameters to determine the spectral imaging sensor design that best take advantage of the MBSIR algorithm.
© (2006) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Travis F. Blake, Matthew E. Goda, Stephen C. Cain, and Kenneth J. Jerkatis "Spectral image improvement analysis of the model-based spectral imaging reconstruction algorithm", Proc. SPIE 6307, Unconventional Imaging II, 630706 (8 September 2006);


Back to Top