[Research of hyperspectral reconstruction based on HJ1A-CCD data]

Area of Science:

• Environmental Remote Sensing
• Water Quality Monitoring
• Hyperspectral Imaging

Background:

• Inland water color remote sensing requires abundant spectral information.
• Hyperspectral imaging offers detailed spectral data crucial for water environment analysis.
• Existing satellite data may have limitations in spectral resolution for precise water monitoring.

Purpose of the Study:

• To reconstruct hyperspectral data from multispectral imagery for improved water environment remote sensing.
• To evaluate the effectiveness of hyperspectral data reconstruction using HJ1A-CCD data.
• To assess the performance of reconstructed data in retrieving water quality parameters like chlorophyll-a concentration.

Main Methods:

• Utilized HJ1A-HSI and HJ1A-CCD satellite imagery acquired on June 6th, 2009.
• Reconstructed hyperspectral data from HJ1A-CCD multispectral imagery.
• Compared reconstructed data with original HJ1A-HSI data using relative error analysis.
• Analyzed image quality metrics including entropy and average gradient.
• Applied a three-band model for chlorophyll-a concentration inversion using both data types.

Main Results:

• The reconstructed hyperspectral data exhibited a lower average relative error (0.3077) compared to the original HJ1A-HSI data (0.3335) in the 660 nm-900 nm range.
• The reconstructed image demonstrated higher entropy and average gradient, indicating improved detail and texture.
• The three-band model achieved higher accuracy in chlorophyll-a concentration inversion when using the reconstructed hyperspectral data.

Conclusions:

• Hyperspectral image reconstruction is a viable method to enhance spectral information for water environment remote sensing.
• Reconstructed data provides a more suitable and accurate data source for inland water color remote sensing applications.
• This technique improves the potential for precise monitoring of water quality parameters from satellite data.