Abstract
PICS (Pseudo Invariant Calibration Sites) have been used for on-orbit radiometric trending of optical satellite sensors for many years. It takes advantage of the properties of PICS which is highly invariant, any trend in the data would indicate a change in sensor responsivity rather than a change in the apparent reflectance of the target/atmosphere system. However, the length of time that is required to determine a change in sensor responsivity is measured by the drift estimates using a function of the residual noise in the PICS target/atmosphere system and the number of days between imaging opportunities with the sensor being calibrated. Often this can require several years of imaging PICS, using only cloud free data, to detect a small change in sensor responsivity. Six primary Saharan PICS locations were selected according to their level of temporal/spatial stability. Each of these sites was normalized to the well-known Libya- 4 PICS which is used as the overall reference calibration site. Pseudo Invariant Calibration sites (PICS) Normalization process (PNP) is a technique developed to combine sensor observations of multiple PICS into a single time series with greater temporal resolution for satellite calibration. As a result, the temporal resolution using this method can theoretically be improved by a factor of four. This PNP technique was applied to Landsat-8 data to determine if small changes in sensor responsivity can be detected in a shorter time period than when only one PICS is utilized in the trending process. It was found that the PNP, using almost 4 years of image observations over 6 primary PICS, can give a sensor trending estimates to within 0.50 % per year for all VNIR and SWIR bands.
Index Terms— PICS, PNP, calibration, trending, Landsat-8
Improved Temporal Resolution of Pseudo Invariant Calibration Sites (PICS) Through Development of the PICS Normalization Process (PNP)
PICS (Pseudo Invariant Calibration Sites) have been used for on-orbit radiometric trending of optical satellite sensors for many years. It takes advantage of the properties of PICS which is highly invariant, any trend in the data would indicate a change in sensor responsivity rather than a change in the apparent reflectance of the target/atmosphere system. However, the length of time that is required to determine a change in sensor responsivity is measured by the drift estimates using a function of the residual noise in the PICS target/atmosphere system and the number of days between imaging opportunities with the sensor being calibrated. Often this can require several years of imaging PICS, using only cloud free data, to detect a small change in sensor responsivity. Six primary Saharan PICS locations were selected according to their level of temporal/spatial stability. Each of these sites was normalized to the well-known Libya- 4 PICS which is used as the overall reference calibration site. Pseudo Invariant Calibration sites (PICS) Normalization process (PNP) is a technique developed to combine sensor observations of multiple PICS into a single time series with greater temporal resolution for satellite calibration. As a result, the temporal resolution using this method can theoretically be improved by a factor of four. This PNP technique was applied to Landsat-8 data to determine if small changes in sensor responsivity can be detected in a shorter time period than when only one PICS is utilized in the trending process. It was found that the PNP, using almost 4 years of image observations over 6 primary PICS, can give a sensor trending estimates to within 0.50 % per year for all VNIR and SWIR bands.
Index Terms— PICS, PNP, calibration, trending, Landsat-8