Abstract
RapidEye is a commercial provider of high-resolution satellite imagery. It operates a constellation of five identical sensors phased equally across a sun-synchronous orbital plane which allows daily revisit times for any place on the globe. Each spacecraft has an optical payload with five spectral bands operating in the visible and near infrared wavelengths. The constellation was launched in August 2008 and became operational in February 2009. The image archive since that time has grown to include over 3 billion square kilometers of data. The RapidEye sensors are pushbroom scanners with 12000 detectors per spectral band. Each satellite, while manufactured to be identical to all the others, is a unique instrument and has had different experiences while being in-orbit. Thus each one must be treated independently in regards to spatial calibration. The difficulties in calculating accurate spatial calibration correction coefficients (relative gains and offsets) for a wide swath width pushbroom camera are well known. The use of both imaging statistics and side slither imagery to provide sufficient detector relative gains and offsets are also well known. These two methods each have their drawbacks though. In order to use imaging statistics, a large number of images must be collected over a long period of time. In a highly competitive field, the time it takes to collect a large enough amount of data can be detrimental to the company. The side slither method allows for relative gains to be calculated from only one image, but only using a single image can introduce artifacts present on the Earth's surface: primarily slightly varying gradients. The planned acquisition of an appropriate side slither calibration site may also interfere with a customer's order in the same area. This paper will detail an approach where the statistical and side slither methods are combined to enhance their respective strengths while limiting their weaknesses in calculating detector relative gains.
Combining Imaging Statistics and Side Slither Imagery to Estimate Relative Detector Gains
RapidEye is a commercial provider of high-resolution satellite imagery. It operates a constellation of five identical sensors phased equally across a sun-synchronous orbital plane which allows daily revisit times for any place on the globe. Each spacecraft has an optical payload with five spectral bands operating in the visible and near infrared wavelengths. The constellation was launched in August 2008 and became operational in February 2009. The image archive since that time has grown to include over 3 billion square kilometers of data. The RapidEye sensors are pushbroom scanners with 12000 detectors per spectral band. Each satellite, while manufactured to be identical to all the others, is a unique instrument and has had different experiences while being in-orbit. Thus each one must be treated independently in regards to spatial calibration. The difficulties in calculating accurate spatial calibration correction coefficients (relative gains and offsets) for a wide swath width pushbroom camera are well known. The use of both imaging statistics and side slither imagery to provide sufficient detector relative gains and offsets are also well known. These two methods each have their drawbacks though. In order to use imaging statistics, a large number of images must be collected over a long period of time. In a highly competitive field, the time it takes to collect a large enough amount of data can be detrimental to the company. The side slither method allows for relative gains to be calculated from only one image, but only using a single image can introduce artifacts present on the Earth's surface: primarily slightly varying gradients. The planned acquisition of an appropriate side slither calibration site may also interfere with a customer's order in the same area. This paper will detail an approach where the statistical and side slither methods are combined to enhance their respective strengths while limiting their weaknesses in calculating detector relative gains.