Spectral-Image-Analysis

This repository hosts tools and resources for analyzing remotely sensed spectral data including Multispectral and Hyperspectral data, covering multiple tasks- from simple wavenumber-to-wavelength domain conversion to complex tasks such as Noise Reduction, Target Detection, Classification and so on. Each directory focuses on a specific task of spectral data processing and analysis and includes all necessary resources.

Current Resources:

• Wavenumber-to-Wavelength domain conversion: Tools for conversion from "Spectral Radiance in Wavenumber [W/cm2 sr cm-1]" to "Spectral Radiance in Wavelength domain [W/m2 sr μm]"

• Spectral-Calibration: Tools for calibrating a Spectrometer using pencil lamps.

• Remotely-Sensed-Spectral-Radiance: Tools for analysis of spectral radiance components and their applications in determining ground-leaving radiance, camouflage detection, and gas plume identification.

• PCA-NAPCA-Noise-Removal: Tools for removing noise from multispectral imagery using Principal Component Analysis (PCA) and Noise Adjusted Principal Component Analysis (NAPC) techniques.

• Atmospheric-Compensation: Tools for performing atmospheric compensation in Hyperspectral imagery using Empirical Line Method (ELM) and generate redlectance data product.

• Hyperspectral-Target-Detection: Tools for detecting targets(Red panel) in Hyperspectral image and performing analysis in both spectral reflectance space and reduced-dimensional space (PCA space).

• Species-Level-Classification-using-HSI-ENVI: An article on using Hyperspectral Imagery to classify forest species using Maximum Likelihood Estimation (MLE) algorithm.

• Land-Degradation-Assessment-using-HSI-ENVI: An article on using Hyperspectral Imagery to assess the status of land degradation using Pixel Unmixing and Spectral Angle Mapper (SAM) classification algorithm.

• Foliar-Biochemistry-Estimation-Minitab: An article on estimating foliar Nitrogen (N) and Potassium (K) concentrations using hyperspectral data and a stepwise linear regression model, evaluating the impact of different spectral processing techniques: raw spectra, first derivative spectra, and continuum-removed spectra.

Note: The ENVI keyword label at the end of directory name indicates that the work has been done using ENVI Application and only analysis reports and results are available.

We welcome suggestions to improve the repository.

Contact: rb1005@rit.edu for potential collaborations.