“…Many other multiwavelength devices and/or processing techniques can be found in the literature: a model to estimate emissivity errors generated during multiwavelength meas urements was developed by Mazikowski and Chrzanowski [94], who concluded that the incorrect selection of the emis sivity function and fluctuations of the temperature of the tar get surface are the largest sources of errors, errors caused by the reflection of background radiation on the target sur face can be quite significant, depending on the difference between background and surface temperatures, but other sources of error, such as electronic noise, calibration inaccu racies, and analogtodigital resolution, can be insignificant in welldesigned pyrometers; Sun et al [95] and Xing and coworkers [96,97] proposed a multiwavelength algorithm, in which, for all wavelengths, a linear relationship between emissivity and true temperature is assumed, so that no spe cific relationship between emissivity and wavelength has to be selected, allowing the simultaneous calculation of true temperatures and spectral emissivities from the measure ment of two consecutive temperatures; Madura et al [98] developed a triplewavelength determined pyrometry sys tem (n = 3 and m = 2) with measuring wavelengths of 1.6, 1.8, and 2 µm, especially suitable for the measurement of temper atures of metallic surfaces, in which the three effec tive radiance temperature outputs and the three possible ratio temperatures (17), estimated from the three dualwavelength combinations, were used to select the most appropriate emis sivity functions; Duvaut [85] developed a multiwavelength pyrometer for temperature and spectral emissivity measure ments of metallic surfaces, applied in the visible range of 0.45-0.7 µm and in the IR range of 2-7 µm, using differ ent emissivity functions, recommending, for temperatures between 600 and 1200 K, systems in the IR rather than the visible range; Svet and Sergeev [99] announced a deter mined triplewavelength pyrometer (n = 3 and m = 2), in which the logarithm of the emissivity is assumed to vary lin early with wavelength, reporting temperature measurement errors below 1%; in multiwavelength pyrometers, since the distance between wavelengths must be low, in order to minimize emissivity variations, but not so low, in order to minimize the uncertainty of temperature estimates, Rodiet and coworkers [100,101] proposed a method for the opti mum selection of wavelengths, through the reduction of the uncertainty of temperature estimates of target surfaces with nonuniform emissivity, considering a multiwavelength determined system with n = 4 spectral channels.…”