Commercial Remote Sensing
- PDF / 2,075,793 Bytes
- 79 Pages / 595.276 x 790.866 pts Page_size
- 25 Downloads / 235 Views
CALIBRATION AND VALIDATION Andreas Colliander Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA
Definition Calibration. The process of quantitatively defining the system responses, under specified conditions, to known, controlled signal inputs. The result of a calibration permits either the assignment of values of measurands to the system output or the determination of corrections with respect to the system output (Joint Committee for Guides in Metrology JCGM (includes ISO) 2008; Randa et al., 2008; CEOS Working Group on Calibration and Validation, 2012). Validation. The process of assessing, by independent means, the quality of the data products derived from the system outputs. The quality is determined with respect to the specified requirements (Joint Committee for Guides in Metrology JCGM (includes ISO) 2008; Randa et al., 2008; CEOS Working Group on Calibration and Validation, 2012). Introduction The value of remotely sensed data products, in the scientific sense in particular, is determined by how well the characteristics of a product are known (e.g., Platt and Sathyendranath, 1988; Wentz and Schabel, 2000; Atlas and Hoffman, 2000; Jung et al., 2010). These characteristics generally include long-and short-term deviation of the product value from the true value corresponding to the measurement, which is estimated through independent means (e.g., Wehr and Attema, 2001), and accuracy of the geographic location assigned to the product
(e.g., Wolfe et al., 2002; Small et al., 2004). The process of determining these characteristics for a particular remote sensing product is referred to as validation. However, before a data product is validated, it needs to be calibrated. Therefore, the calibration and validation processes are very closely linked together although they are distinctively two separate processes (see the Definition). This entry discusses calibration and validation in terms of characterization against the true value; geolocation aspect of the validation is a separate topic with specific challenges and solutions. Remote sensing missions have requirements for the data products they are tasked to produce (e.g., Barre et al., 2008). The aim of the calibration and validation process of a particular mission is then to show that it meets its stated requirements (e.g., Delwart et al., 2008). Since the requirements are typically assigned based on expected scientific utilization of the data, the calibration and validation processes are generally regarded as a scientific function. Furthermore, the science community commonly contributes to calibration and validation efforts of data products independently from the missions in their research, due to the importance of knowing the characteristics and quality of the data (e.g., Donlon et al., 2002; Wang and Key, 2003; Mears and Wentz, 2005; Flanner et al., 2010). The challenges of calibration and validation are specific to the mission and the data product. However, there are some general challenges concerning most of the remote sensing pr
Data Loading...
