Implementation of Stochastic Modelling in Enhanced Cadastral Databased for Multi-Classes Datasets

Abstract

Stochastic Modelling (SM) was a crucial component of least squares adjustment (LSA), particularly when processing data from geodetic networks. The projected variances which generate using SM execute an important part in defining both the accurateness of the computed parameter vectors and the impact of the adjustment outcomes. As positional precision becomes the primary objective, there is still potential for improvement because there are multiple sources of datasets with varying levels of data quality. Concerning the assertion that the National Digital Cadastral Database (NDCDB) is accurate, its development involved the use of historical datasets that were obtained from a number of different measurement classes, specifically the first, second, and third classes. In this study, researchers evaluated whether or not it is possible to employ stochastic modelling to maintain the position correctness of historical data that encompasses a wide range of data quality classes. In order to accomplish this, an approach known as an Least Squares Variance Component Estimator (LS-VCE) was utilised to generate reliable estimates of variances. Two (2) certified plans (CPs) that is CP93887 and CP33758 was selected as measurements for the first and second classes CP, respectively. The experiment showed that the variance that has been estimated by LS-VCE could produce realistic adjustment results, as shown by an analysis of the corrected results obtained by allocating the variance into different data classes. In light of these findings, the investigations showed and demonstrated conclusively that separate variance is necessary for each data classes with the aim of preserving positional accuracy. In conclusion, it is crucial to incorporate a realistic variance component inside a coordinated cadastral database in order to fulfil the objective of ensuring the accurateness of survey data for future time periods.