In pyrometry measurements, the unknown target emissivity is a critical source of uncertainty, especially when the emissivity is low. Aiming to overcome this problem, various multi-spectral pyrometry systems and processing techniques have been proposed in the literature. Basically, all multi-spectral systems are based on the same principle: the radiation emitted by the target is measured at different channels having different spectral characteristics, and the emissivity is modelled as a function of wavelength with adjustable parameters to be obtained empirically, resulting in a system of equations whose solution is the target temperature and the parameters of the emissivity function. The present work reviews the most important multi-spectral developments. Concerning the spectral width of the measurement channels, multi-spectral systems are divided into multi-wavelength (monochromatic channels) and multi-band (wide-band channels) systems. Regarding the number of unknowns and equations (one equation per channel), pyrometry systems can either be determined (same number of unknowns and equations, having a unique solution) or overdetermined (more equations than unknowns, to be solved by least-squares). Generally, higher-order multi-spectral systems are overdetermined, since the uncertainty of the solutions obtained from determined systems increases as the number of channels increases, so that determined systems normally have less than four channels. In terms of the spectral characteristics of the measurement channels, narrow bands, far apart from each other and shifted towards lower wavelengths, seem to provide more accurate solutions. Many processing techniques have been proposed, but they strongly rely on the relationship between emissivity and wavelength, which is, in turn, strongly dependent on the characteristics of a particular target. Several accurate temperature and/or emissivity results have been reported, but no universally accepted multi-spectral technique has yet been developed.

Condition monitoring is a fundamental part of machining, as well as other manufacturing processes where, generally, there are parts that wear out and have to be replaced. Devising proper condition monitoring has been a concern of many researchers, but there is still a lack of robustness and efficiency, most often hindered by the system’s complexity or otherwise limited by the inherent noisy signals, a characteristic of industrial processes. The vast majority of condition monitoring approaches do not take into account the temporal sequence when modelling and hence lose an intrinsic part of the context of an actual time-dependent process, fundamental to processes such as cutting. The proposed system uses a multisensory approach to gather information from the cutting process, which is then modelled by a recurrent neural network, capturing the evolutive pattern of wear over time. The system was tested with realistic cutting conditions, and the results show great effectiveness and accuracy with just a few cutting tests. The use of recurrent neural networks demonstrates the potential of such an approach for other time-dependent industrial processes under noisy conditions.