Two kinds of time records are found when measuring electrochemical noise: those presenting characteristic transients or spikes, and others in which neither transients nor spikes can be distinguished. Although the former are likely to be more informative than the latter, most of the methods developed for analyzing electrochemical noise are focused on studying the latter. This paper suggests a methodology for automatically studying the appearance of the transient, which could represent a significant advance in the field of corrosion monitoring and laboratory investigation. The proposed methodology is based on a mathematical tool called the wavelet transform, whose principal characteristic is that it enables a simultaneous analysis of signals in both time and scale. Thus a subsequent interscalar analysis of the signals can reveal the existence of transients and even provide a comparative characterization of their size and scale.