New line accuracy assessment methodology using nonlinear least-squares estimation

This paper presents a novel line accuracy assessment technique by measuring the offsets of a typical measured line from the true reference line. These measurements are assumed to follow a Gaussian distribution. Buffers of gradually increasing widths, drawn around the true line, are used to measure the magnitudes of line offsets from true locations. A nonlinear least-squares estimation is used to determine the mean and the standard deviation of the line offset. The purpose of the proposed parameter estimation technique is to improve, using the two Gaussian parameters of mean and standard deviation, the line error modeling, and to uncover the physical meaning of the magnitude and variability of line offsets, respectively. The feasibility of the parameter estimation technique is demonstrated by a series of tests that confirm the assumption of a nonzero mean Gaussian distribution. The proposed methodology is expected to provide better insight into the spatial data quality of linear features in geographical information systems.