Back to results list
Please use this identifier to cite or link to this item:
|Title:||Reliability of spatial data and its analysis in GIS||Authors:||Cheung, Chui-kwan||Keywords:||Geographic information systems.
Hong Kong Polytechnic University -- Dissertations
|Issue Date:||2000||Publisher:||The Hong Kong Polytechnic University||Abstract:||Geographical data in geographical information system (GIS) are not error-free. Accuracy of each object in the GIS should be attached with their data description. This is particularly important when the data is used for decision-making. In this study, we focus on modeling positional error of spatial features in GIS. A reliability model of a spatial feature is proposed in this study. It is measured by discrepancies of the spatial feature. Since the measured spatial feature may contain positional errors, a simulation technique is adopted to simulate the positional errors of the spatial features. Most possible measured spatial features are generated based on the assumption of the nodal errors' distribution. For each measured spatial feature generated in the simulation, its discrepant area can be computed. An average of discrepant areas is an indicator of the reliability of the spatial features. In this study, we describe three further developments on the reliability of line segment, which is a basic unit of the linear feature of GIS, to the previous studies. First, two possible statistical distributions, both uniform and bivariate normal distributions, of the errors of line segment's nodes are discussed. While in the previous studies, the uniform distribution was the only distribution case discussed. Second, an error ellipse model, instead of the error circle model, is used for describing the errors of the nodes. Third, an effect of error dependent relationship of two nodes on the reliability of line segment is further discussed. From our results, it is noticed that different combinations of correlated nodal errors yield different reliability of a line segment.
Apart from the simulation approaches, another reliability model is derived from a newly developed approach - the numerical integration technique - in order to validate the simulated results, mainly due to the fact that accuracy of the simulation approaches has caused worry in some circles. After comparing there two methods, we notice that simulated and the numerical results are approximately the same, but they have different computational time. In the reliability model of a line segment, we can achieve the numerical result in a shorter time. On the other hand, the simulated result can be obtained in a shorter time in other reliability models. This is due to the complexity of the numerical model depending on the amount of nodes of the spatial feature itself. The reliability model based on both methods is extended to the reliability of both 2D and 3D linear features, both 2D and 3D areal features, and 3D volumetric features in GIS. It also concludes that error ellipse parameters affect the reliability of a spatial feature. Furthermore, an error propagation model in buffer spatial analysis is derived based on both the simulation and the numerical analysis approaches. It is observed that the size of a buffer affects the reliability of the buffer. The reliability model proposed in this project is thus applicable to all features of GIS and buffer GIS operations for error description.
|Description:||[ix], 109 leaves : ill. ; 30 cm.
PolyU Library Call No.: [THS] LG51 .H577M LSGI 2000 Cheung
|URI:||http://hdl.handle.net/10397/3581||Rights:||All rights reserved.|
|Appears in Collections:||Thesis|
Show full item record
Files in This Item:
|b15177853_link.htm||For PolyU Users||162 B||HTML||View/Open|
|b15177853_ir.pdf||For All Users (Non-printable)||3.58 MB||Adobe PDF||View/Open|
Citations as of Feb 12, 2018
Citations as of Feb 12, 2018
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.