" In the early days of PIV, the images suffer from a low signal-to-noise ratio because of the low intensity of available lasers and the poor sensitivity of the recording media. In combination with an inhomogeneous seeding distribution, this resulted in large amounts of spurious vectors at that time. Consequently, the attention of the researchers was mainly directed toward the problem of false vector detection (Keane and Adrian 1990, 1992). In the following decade, the quality of the PIV components has strongly improved, and instead of spurious vectors, the uncertainty of valid vectors was of major interest. This stimulated the development of advanced evaluation techniques which became possible due to the advances in computer power (Stanislas et al. 2003, 2005, 2008). In recent years, another transition can be observed toward the reliable quantifcation of the PIV uncertainty. In order to compare the mean velocity distributions or higher-order statistics from experiments and numerical predictions, for instance, it is essential to quantify the uncertainty of the estimated values. However, due to the complex recording and evaluation procedure of PIV, the error estimation is quite challenging." "Many parameters, including particle image size, particle image density, turbulent fluctuations, noise level, velocity gradients and interrogation window size, affect the uncertainty. In the last years, different methods were developed to quantify the uncertainty of PIV velocity felds (Charonko and Vlachos 2013; Kähler et al. 2012b; Neal et al. 2015; Sciacchitano et al. 2013; Timmins et al. 2012; Wieneke 2015; Wilson and Smith 2013; Xue et al. 2015; Sciacchitano et al. 2015; Christensen and Scarano 2015). Two promising strategies have emerged: The frst one is based on identifying all parameters that influence the measurement uncertainty and determining their effect on the overall uncertainty (Wilson and Smith 2013). This requires that all relevant parameters and sensitivities are known. The second approach reduces the parameter space by analyzing the correlation functions only. This is motivated by the fact that the shape of the correlation function is a result of all parameters that contribute to the measurement uncertainty (Charonko and Vlachos 2013; Wieneke 2015)."
% Fontes ruidosas - Veja "Current limitations" da Pag 5 da Tese de Theunissen (2010). - O efeito do ruído de fundo na qualidade da imagem é apresentado nas página 83 e 84 do trabalho de Sciacchitano (2014).
