Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Weighted Sum of Segmented Correlation: An Efficient Method for Spectra Matching in Hyperspectral Images

Sampriti Soor, Priyanka Kumari, B. S. Daya Sagar, Amba Shetty

TL;DR

The paper tackles material identification in hyperspectral imagery by arguing that absorption-feature positions and shapes offer the most discriminative information. It introduces Weighted Sum of Segmented Correlation (WSSC), which computes segment-wise correlation indices and combines them into a weighted matching index, using normalization to zero mean and unit variance and a Pearson-like correlation for each segment. Experimental results on AVIRIS Cuprite (Earth) and CRISM MTRDR (Mars) show that WSSC outperforms full-spectrum cosine similarity and correlation coefficient by leveraging informative spectral segments, enabling accurate identification of minerals such as alunite, kaolinite, chalcedony, Fe olivine, and Mg-smectites even in mixed or subtle signatures. The method offers a practical and scalable approach for hyperspectral spectrum matching with potential applications in remote sensing of planetary surfaces and mineral mapping.

Abstract

Matching a target spectrum with known spectra in a spectral library is a common method for material identification in hyperspectral imaging research. Hyperspectral spectra exhibit precise absorption features across different wavelength segments, and the unique shapes and positions of these absorptions create distinct spectral signatures for each material, aiding in their identification. Therefore, only the specific positions can be considered for material identification. This study introduces the Weighted Sum of Segmented Correlation method, which calculates correlation indices between various segments of a library and a test spectrum, and derives a matching index, favoring positive correlations and penalizing negative correlations using assigned weights. The effectiveness of this approach is evaluated for mineral identification in hyperspectral images from both Earth and Martian surfaces.

Weighted Sum of Segmented Correlation: An Efficient Method for Spectra Matching in Hyperspectral Images

TL;DR

The paper tackles material identification in hyperspectral imagery by arguing that absorption-feature positions and shapes offer the most discriminative information. It introduces Weighted Sum of Segmented Correlation (WSSC), which computes segment-wise correlation indices and combines them into a weighted matching index, using normalization to zero mean and unit variance and a Pearson-like correlation for each segment. Experimental results on AVIRIS Cuprite (Earth) and CRISM MTRDR (Mars) show that WSSC outperforms full-spectrum cosine similarity and correlation coefficient by leveraging informative spectral segments, enabling accurate identification of minerals such as alunite, kaolinite, chalcedony, Fe olivine, and Mg-smectites even in mixed or subtle signatures. The method offers a practical and scalable approach for hyperspectral spectrum matching with potential applications in remote sensing of planetary surfaces and mineral mapping.

Abstract

Matching a target spectrum with known spectra in a spectral library is a common method for material identification in hyperspectral imaging research. Hyperspectral spectra exhibit precise absorption features across different wavelength segments, and the unique shapes and positions of these absorptions create distinct spectral signatures for each material, aiding in their identification. Therefore, only the specific positions can be considered for material identification. This study introduces the Weighted Sum of Segmented Correlation method, which calculates correlation indices between various segments of a library and a test spectrum, and derives a matching index, favoring positive correlations and penalizing negative correlations using assigned weights. The effectiveness of this approach is evaluated for mineral identification in hyperspectral images from both Earth and Martian surfaces.
Paper Structure (6 sections, 9 equations, 3 figures)

This paper contains 6 sections, 9 equations, 3 figures.

Table of Contents

  1. Introduction
  2. Weighted Sum of Segmented Correlation
  3. Application on Mineral Identification
  4. AVIRIS data green1998imaging (Earth surface)
  5. CRISM data nasapds (Martian Surface)
  6. Conclusion

Figures (3)

  • Figure 1: Top Row: A preprocessing pipeline translates the two H2O-Ice spectra (from MICA spectral library viviano2014revised and a CRISM MTRDR data) in the left image to the right image. Middle Row: Band-minima and FWHMs for all the prominent segments in the processed spectra. Bottom Row: Correlation coefficients ($c^W$) of the spectra sgments and the matching index ($I$) between the two preprocessed spectra.
  • Figure 2: Some dominant minerals in the Alunite hill scene from AVIRIS Cuprite data. Ground-truth obtained from clark2003imaging.
  • Figure 3: Dominant minerals in CRISM MTRDR data FRT3E12 from Nili Fossae region of Mars. In CAR browse product the distribution of Mg-Smectites is indicated in magenta, while in MAF the presence of Fe Olivines is indicated in red.