Abstract
An important part of any earth observing sensor’s calibration is determining the precise location of the sensor footprint on the earth (geolocation calibration). SDL has evaluated the geolocation calibration of two sensors on the Suomi-NPP spacecraft: ATMS (Advanced Technology Microwave Sounder) and CrIS (The Cross-track Infrared Sounder). Both sensors have bands that can sense the radiance contrasts that generally occurs at land sea boundaries both during the day and the night. The observed position of the land sea transitions can be compared to the know shoreline position to evaluate the geolocation calibration. SDL used two different methods to achieve sub-pixel (footprint) resolution. The first method used a cubic polynomial fit to four sensor measurement pixels that crossed the land sea boundary. The shoreline position was then calculated as the inflection point of the polynomial. The second method created a number of synthetic scenes with shifted map positions to find the best fit to the observed scene. The fraction of land to sea in each footprint was used to calculate the simulated scene radiances. The two methods are quite complementary. The inflection point method is generally simpler and does not require a detailed knowledge of the sensor footprint. While the land fraction method is less effected by irregular shoreline structures or irregular sensor footprint spacing. Both methods were computationally demanding and were greatly benefited by software optimization. Results for both ATMS and CrIS geolocation will be presented.
ATMS and CrIS Geolocation
An important part of any earth observing sensor’s calibration is determining the precise location of the sensor footprint on the earth (geolocation calibration). SDL has evaluated the geolocation calibration of two sensors on the Suomi-NPP spacecraft: ATMS (Advanced Technology Microwave Sounder) and CrIS (The Cross-track Infrared Sounder). Both sensors have bands that can sense the radiance contrasts that generally occurs at land sea boundaries both during the day and the night. The observed position of the land sea transitions can be compared to the know shoreline position to evaluate the geolocation calibration. SDL used two different methods to achieve sub-pixel (footprint) resolution. The first method used a cubic polynomial fit to four sensor measurement pixels that crossed the land sea boundary. The shoreline position was then calculated as the inflection point of the polynomial. The second method created a number of synthetic scenes with shifted map positions to find the best fit to the observed scene. The fraction of land to sea in each footprint was used to calculate the simulated scene radiances. The two methods are quite complementary. The inflection point method is generally simpler and does not require a detailed knowledge of the sensor footprint. While the land fraction method is less effected by irregular shoreline structures or irregular sensor footprint spacing. Both methods were computationally demanding and were greatly benefited by software optimization. Results for both ATMS and CrIS geolocation will be presented.