|
There are many different types of radiometric measurements. They may be classified as measurements of flux in its many different geometric forms; measurements of the material properties of transmittance, reflectance, emissivity, absorptance, and scattering; and properties of instruments. Other aspects of measurement include the restrictions and effects of the different independent variables: spectrum, time variation, angle, position, and extent. Approaches to making the measurements and accounting for the extraneous and intrinsic influences are given, and some of the applications are contained in this chapter.
11.1 Relative and Absolute Measurements
Absolute measurements are always harder to make than relative ones, like absolute truth is harder than relative truth (honor and integrity)! A relative measurement may be considered to be the ratio of output signal levels. It may be a set of signals compared to the first one, as in a relative spectral transmission. The value of the transmission is never determined, only how it changes from one part of the spectrum to another. An absolute measurement requires a calibration so that the value of the radiometric quantity can be found. For this example it would be the transmittance of the sample, perhaps at all wavelengths of interest. Sometimes these can be combined: an absolute measurement is made at one wavelength, and relative measurements are made with respect to this value.
11.2 Errors
The quality of a measurement is usually described in terms of its accuracy and precision. Actually, most people state the inaccuracy and imprecision since low numbers are cited. Typically one might cite an accuracy of 1%, actually meaning an uncertainty in the accuracy of 1%, but that is tradition and shortspeak.
Accuracy is the degree to which the actual value is measured. The uncertainty in accuracy is a value that represents how close to truth the measurement probably comes. Precision is somewhat easier to grasp. It is the repeatability of a measurement. The uncertainty in accuracy can be no less than the variability, the lack of repeatability, the (im)precision of the measurement.
|
