Positional accuracy improvement: lessons learned from regional Victoria, Australia

Abstract

Many spatial datasets in use today have been manually digitised from hard-copy maps. These datasets are typically of relatively low positional accuracy compared to the latest positioning technologies, such as those based on Global Navigation Satellite Systems (GNSS). Discrepancies between the legacy dataset and GNSS positioning solutions are becoming increasingly evident, a problem that is compounded by the growing trend to underlay such datasets with high resolution imagery. As a result, many data providers are seeking methods to update their spatial datasets.In the example of the Victorian land parcel map base, survey-accurate data could be used to improve the positional accuracy of the existing cadastral dataset. However, since the higher accuracy data is only available in localised pockets, its integration into the map base will require rigorous methods that also enable the resultant spatial variation in the quality of the upgraded dataset to be communicated. Adjustment software based on the method of least squares is able to provide optimal positioning solutions that take into account all of the available information, including geometric constraints and the quality of the input datasets. Moreover, it provides updated quality parameters at the individual coordinate level.A case study is presented here, demonstrating the use of such adjustment software to improve the positional accuracy of a sub-set of the Victorian cadastral database using GNSS data from an extensive survey. Initial methods to communicate the spatial variation in the quality of the updated dataset are also developed. Three separate adjustments were performed, with increasingly sophisticated data inputs. An independent ground survey was then undertaken to validate the results reported by the software. The findings indicate that the reported adjustment results are dependent upon the correctness of assumptions made in its establishment. It can be concluded that adjustment software based on the method of least squares is appropriate for positional accuracy improvement of spatial datasets only if sufficient care is given to verifying any assumptions made.