The General Motors rapid travel profilometer has been evaluated by the Michigan Department of State Highways. It meets or exceeds all specifications for accuracy and reliability. It does not return a survey type of elevation map because long-wave features must be filtered out. For this reason, its profiles must be viewed as correct in the frequency domain but incorrect in the spatial domain. An inertial guidance system capable of recording long-wave features would solve the problem. Profile analysis in the frequency domain is confined to the 4 basic measures: mean squares, amplitude distributions, autocorrelation, and power spectral density. It is possible to extract some single number indexes based on the 4 standard measures. Power spectral density appears to be most interesting for highway work. /author/
Abstract