Removing Cosmic Spikes Using a Hyperspectral Upper-Bound Spectrum Method.

Area of Science:

• Spectroscopy
• Chemometrics
• Image Analysis

Background:

• Cosmic ray spikes are a significant challenge in hyperspectral imaging, often corrupting spectral data and hindering chemometric analysis.
• The large volume of spectral data in hyperspectral imaging presents both a challenge (more spikes) and an opportunity (better detection).

Purpose of the Study:

• To develop an improved method for cosmic ray spike removal in hyperspectral images.
• To evaluate the performance and spectral bias introduced by the new method compared to existing techniques.

Main Methods:

• Leveraging the spatial dimension of hyperspectral data matrices (DM) with principal component analysis (PCA).
• Expanding the Upper Bound Spectrum-DM (UBS-DM) method into a new hyperspectral-specific approach (UBS-DM-HS).
• Identifying cosmic spikes by analyzing principal components present in limited pixels.

Main Results:

• The developed UBS-DM-HS method demonstrates effective removal of cosmic ray spikes from hyperspectral data.
• Comparison with other despiking methods shows the efficacy of UBS-DM-HS in preserving spectral integrity.
• The method was validated using both simulated and experimental hyperspectral Raman image datasets.

Conclusions:

• The UBS-DM-HS method offers a robust solution for cosmic ray spike removal in hyperspectral imaging.
• This advancement improves the reliability of chemometric analysis on hyperspectral datasets.
• The method effectively balances spike removal with minimal spectral bias.