| Aerosol Optical Depth(AOD)is a crucial parameter for determining the radiative effects of aerosols and the state of air quality.Detecting AOD at nighttime is highly valuable for understanding the nocturnal distribution of aerosols,analyzing their diurnal variations,and predicting atmospheric pollution.However,quantitatively retrieving nighttime AOD using passive satellite remote sensing remains a significant challenge due to the absence of a stable light source(such as the sun)during nighttime hours.Currently,there is a lack of available nighttime AOD products.Low-light satellites have the capability to detect natural light sources(such as moonlight,firelight,etc)and artificial light sources on the Earth’s surface during nighttime,providing unique information about nocturnal Earth phenomena.Previous studies have utilized NPP/VIIRS low-light imagery to quantify nighttime AOD,but the reliability of the retrieved results and the spatial resolution are limited.The SDGSAT-1 satellite,equipped with a10-meter panchromatic low-light band and a 40-meter RGB low-light band,offers a new source of low-light data that is expected to yield more refined results for nighttime AOD retrieval.However,applying existing nighttime AOD retrieval algorithms to the high-resolution SDGSAT-1 low-light data presents several challenges that need to be assessed and resolved,owing to disparities in spatial resolution,spectral response,and image quality between the two sensors.Therefore,this study focuses on the assessment of the quality of SDGSAT-1 low-light data,noise reduction techniques,and the quantitative retrieval of nighttime aerosol optical depth.A denoising algorithm has been developed,which,when applied to SDGSAT-1low-light data,produces refined results for nighttime AOD retrieval.The primary research content and achievements are as follows:(1)The dynamic range and signal-to-noise ratio of the SDGSAT-1 low-light imager were evaluated,and the results exceeded the design values,indicating that the sensor performance meets the design requirements.Through a comparative analysis of radiometric performance with NPP/VIIRS data,the correlation coefficient(R~2)between the two was determined to be0.77,indicating that the radiometric calibration of SDGSAT-1 low-light data is relatively reliable,allowing for extensive quantitative applications.(2)To address the issue of striping noise in SDGSAT-1 low-light data,a simple and efficient spectral-correlation denoising(SCD)algorithm was proposed in this study.The algorithm effectively removes noise while preserving essential information.Comparative results with two other image denoising methods demonstrate that the SCD algorithm offers superior denoising performance and robustness.Additionally,the algorithm minimizes the impact of moonlight on SDGSAT-1 low-light data,thus enhancing data comparability across different lunar phases.(3)Beijing is chosen as the study area,and using SDGSAT-1 low-light data,an improved method for synthesizing multi-temporal background data was developed.Two methods,the Radiative Background(RB)method and the Standard Deviation(SD)method,were employed for nighttime AOD retrieval.The results indicate a high consistency between the retrieved nighttime AOD values obtained using both methods and the measured data from AERONET sites.When utilizing SDGSAT-1 low-light data for AOD retrieval,the RB method outperforms the SD method.Finally,a brief analysis of the error sources in the retrieval results is presented.The study confirms that SDGSAT-1 can yield relatively reliable nighttime AOD retrieval,thereby providing a foundation for generating scientific products related to nighttime AOD. |
