[A New Method of Accurately Extracting Spectral Values for Discrete Sampling Points]

Area of Science:

• Remote Sensing
• Geospatial Analysis
• Image Processing

Context:

• Remote sensing information inversion models rely on accurate spectral data from discrete sampling points.
• Existing methods of extracting spectral values using regions of interest (ROIs) in ENVI software can lead to coordinate inconsistencies and inaccurate spectral data.
• These inaccuracies compromise the validity of remote sensing inversion models.

Purpose:

• To systematically study the principles of extracting target coordinates and spectral values.
• To develop and validate a new, accurate method for extracting spectral parameters from discrete sampling points within ENVI software.
• To address the limitations of existing spectral extraction techniques.

Summary:

• The study identifies inaccuracies in spectral value extraction when using original coordinates in ENVI, particularly when sampling points do not align with the upper-left corner of a pixel.
• A novel method is proposed that adjusts sampling point coordinates to the upper-left corner of pixels using symmetry principles before spectral extraction.
• Experiments on OLI, TM, and ETM+ imagery confirm the accuracy and feasibility of the new method for extracting discrete spectral values.

Impact:

• Provides a more reliable method for spectral data extraction, enhancing the accuracy of remote sensing inversion models.
• Offers a clear, simple, and broadly applicable solution for researchers working with discrete spectral data.
• Improves the fundamental data quality for various remote sensing applications.