This article describes a procedure for adjusting sampling locations for inspected datasets of track geometry. These sets are spatially discretized and their sampling locations are never reproduced. Owing to these location gaps, the variation of track geometry could not be computed from the differences between two datasets. The main difficulty of the adjustment is derived from the fact that some sampling intervals locally lengthen and shorten. In this procedure, the adjustment is formulated into an optimization form and unknown parameters involved in the form are statistically identified. Consequently, the variation datesets are obtained and they demonstrate new properties.

Copyright (c) 2007 Railway Technical Research Institute