Abstract
It is important for a sensor to provide calibrated, consistent measurement of the Earth’s surface feature through a long period of time. HJ-1A, launched on Sept. 6, 2008, is the first satellite of the Environment and Disasters Monitoring Microsatellite Constellation of China. This work focuses on monitoring the long-term on-orbit radiometric stability of the HJ-1A CCD sensors using Dunhuang Site. This site is one of the national calibration sites of China and has been frequently used as radiometric calibration target because of its relatively stable surface conditions spatially and temporally. This study was performed using all cloud-free images from the HJ-1A CCD sensors and the Terra MODIS, acquired from HJ-1A’s launch to August 2013. The Terra MODIS was take as the radiometric reference and the cross-calibration method was used to derive the calibration coefficients of the HJ-1A CCD sensors. Homogeneous regions of interest (ROI) were selected in the HJ-1A CCD and Terra MODIS images pairs and the mean statistics were derived from the HJ-1A CCD sensors in terms of DNs and from the Terra MODIS in terms of top-of-atmosphere (TOA) reflectance respectively. To reduce the error of the cross-calibration results of the HJ-1A CCD sensors, the differences due to the Relative Spectral Response (RSR) characteristics of the HJ-1A CCD sensors and Terra MODIS and the bidirectional reflectance distribution function (BRDF) of the calibration target were considered. Based on the cross-calibration results, the radiometric stability of HJ-1A CCD sensors was assessed. The results show that the radiometric performance of each band of the HJ-1A CCD sensors almost has the same tendency of degradation. But the rates of the degradation of the four bands of HJ-1A CCD sensors are obviously different. The blue band has the biggest degradation, the relative error of the calibration coefficients during the six years is more 20%. The infrared band has the smallest degradation and the relative error is no more than 10%.
Monitoring On-orbit Radiometric Stability of HJ-1A/CCD using TERRA/MODIS Data
It is important for a sensor to provide calibrated, consistent measurement of the Earth’s surface feature through a long period of time. HJ-1A, launched on Sept. 6, 2008, is the first satellite of the Environment and Disasters Monitoring Microsatellite Constellation of China. This work focuses on monitoring the long-term on-orbit radiometric stability of the HJ-1A CCD sensors using Dunhuang Site. This site is one of the national calibration sites of China and has been frequently used as radiometric calibration target because of its relatively stable surface conditions spatially and temporally. This study was performed using all cloud-free images from the HJ-1A CCD sensors and the Terra MODIS, acquired from HJ-1A’s launch to August 2013. The Terra MODIS was take as the radiometric reference and the cross-calibration method was used to derive the calibration coefficients of the HJ-1A CCD sensors. Homogeneous regions of interest (ROI) were selected in the HJ-1A CCD and Terra MODIS images pairs and the mean statistics were derived from the HJ-1A CCD sensors in terms of DNs and from the Terra MODIS in terms of top-of-atmosphere (TOA) reflectance respectively. To reduce the error of the cross-calibration results of the HJ-1A CCD sensors, the differences due to the Relative Spectral Response (RSR) characteristics of the HJ-1A CCD sensors and Terra MODIS and the bidirectional reflectance distribution function (BRDF) of the calibration target were considered. Based on the cross-calibration results, the radiometric stability of HJ-1A CCD sensors was assessed. The results show that the radiometric performance of each band of the HJ-1A CCD sensors almost has the same tendency of degradation. But the rates of the degradation of the four bands of HJ-1A CCD sensors are obviously different. The blue band has the biggest degradation, the relative error of the calibration coefficients during the six years is more 20%. The infrared band has the smallest degradation and the relative error is no more than 10%.