Accuracy Conversion Parameters Other Inputs and Outputs

“Service Selection” inputs whose data type is an enumeration of all the alternative specific services available through one service operation. These “Service Selection” inputs could be handled like, or as an extension of, the Strategy Parameters discussed above.

4.12. Other Inputs and Outputs

Some image exploitation service inputs and outputs do not fall in the above categories. Table 4-2 lists many of these other inputs and outputs, with some information on the possible data format, mostly using ISO standard IDL data types and structures. Service Input or Output Possible Data Format Computed dimensions, including: Distance, Area, Volume Angle Time Difference Velocity NameValueList See Section 4.2, or use: typedef float Distance; Units: meters typedef float Area; Units: meters squared typedef float Volume; Units: meters cubed typedef float Angle; Units: degrees typedef float TimeDifference; Units: seconds typedef float Velocity; Units: meters per second Dimension accuracy estimates typedef float Variance; Units: Dimension units squared Transformation accuracy estimates See Position Accuracy Estimates, Section 3.5 Accuracy estimates See Position Accuracy Estimates, Section 3.5, and: typedef float CE; Units: meters typedef float LE; Units: meters typedef float SphericalError; Units: meters Confidence level for CE and LE typedef float Confidence; Units: probability Accuracy conversion parameters NameValueList See Section 3.12.1 Measures of quality of image match typedef float Confidence; Units: probability See Position Accuracy Estimates, Section 3.5 Success or failure of metadata modifications typedef boolean OperationSucceeded; Table 4-2. Possible Formats of Other Service Inputs and Outputs Note that all quantities listed as “float” type could alternately be “double” type, if more accuracy is needed. The quantities listed as “float” type are considered unlikely to need more accuracy than provided by “float”.

4.12.1. Accuracy Conversion Parameters

The “Accuracy conversion parameters” might be formatted as a Name Value List, as discussed below in Section 4.14. For conversion between covariance matrices and CE plus LE, the names and values shown in Table 4-3 might be used: Name Data Type Units Definition Comments Error Probability float none Confidence probability for CE and LE error estimates Note A Probability Distribution Name string none Name of probability distribution, such as “normal distribution” LE Multiplier float none Multiplier of standard deviation to obtain LE with specified probability Note B CE Multiplier Function sequence float none Multipliers of standard deviation to obtain CE with specified probability Note C Table 4-3. Possible Accuracy Conversion Parameters Table Notes: The Open GIS Abstract Specification Page 49 Volume 15: Image Exploitation Services 00-115 corrigendum A. The possible probability values include 0.50, 0.6827, 0.90, and 0.95. B. This multiplier is used to compute LE with the specified confidence probability from the standard deviation of the error square root of the variance. C. These multipliers are used to compute CE with the specified confidence probability from the standard deviation square root of the variance along the major axis of the error ellipse. The CE is computed from the covariance matrix of the expected errors in two axes, either the two horizontal ground coordinates or the two image coordinates. When the error estimates in the two axes are correlated andor are not equal, the variances are first computed for the major and minor axes of the error ellipse. The ratio of the variance in the minor axis to the variance in the major axis of the error ellipse is then computed. As this ratio varies from 0.0 to 1.0, the multiplier needed to compute CE varies from one number to a larger number. The multipliers are specified for a variable number of evenly spaced values of the minormajor axis variance ratio, from 0.0 to 1.0. The number of ratio values may be 21, for ratio values differing by 0.05. For a minormajor ratio between the recorded values, linear interpolation is used.

4.13. Other Possible Outputs