[Evaluation of four dark object atmospheric correction methods based on ZY-3 CCD data].

Area of Science:

• Remote Sensing
• Atmospheric Correction
• Image Processing

Context:

• Evaluation of dark-object subtraction (DOS) atmospheric correction methods.
• Utilized 2012 Inner Mongolia experimental data.
• Applied to ZY-3 CCD satellite imagery.

Purpose:

• To assess the performance and accuracy of four different DOS atmospheric correction algorithms.
• To analyze the impact of key parameters on DOS method effectiveness for ZY-3 data.
• To identify the optimal DOS method for atmospheric correction of ZY-3 imagery.

Summary:

• All four DOS methods showed significant atmospheric correction effects on bands 1, 2, and 3.
• DOS4 exhibited the best performance on band 4, while DOS2 performed the worst.
• DOS1 and DOS3 showed limited atmospheric correction effects, with DOS1 having >10% relative error across bands.

Impact:

• The DOS4 method achieved the highest accuracy, with absolute error <0.02 and relative error <10% across all bands.
• DOS2 showed the best results on band 1 (RE=4.32%) but the worst on band 4 (RE=19.12%).
• DOS3 consistently yielded a relative error of approximately 10% for all bands.