Spatial Integrity Constraints: A Tool for Improving the Internal Quality of Spatial Data, Sylvain Vallières, Jean Brodeur and Daniel Pilon. Uncertainty in boundaries of interpretive polygonal maps; 5. None of these representations corresponds strictly to reality, but these models represent the same reality at different levels of abstraction, to meet different needs. Examples of poorly defined geographical objects 19 3. If several people were asked to name the best brand of car, the best book, or the best cake, there would be a wide range of different answers.
In this model, the three categories of source errors create primary errors positional errors and attribute errors and secondary errors logical consistency and completeness. Using expert knowledge to reason with uncertainty. On the Importance of External Data Quality in Civil Law Marc Gervais. The evaluation of 38 Fundamentals of Spatial Data Quality internal quality includes an external part comparison with reference data , as well as an internal part, depending on the elements of quality to be verified. Overview of the radiometric defects; 4. The Impact of Positional Accuracy on the Computation of Cost Functions.
In the same way, a set of data of great internal quality representing the hydrology of a region, created by following the best quality control, could be very useful that is, good external quality to an environmental expert and almost useless that is, poor external quality to a land surveyor. All of these errors form the global error attached to the final product. Abstract: This book explains the concept of spatial data quality, a key theory for minimizing the risks of data misuse in a specific decision-making context. Research into probability and statistics, artificial intelligence, and databases have examined the question of uncertainty for many years. To evaluate whether a dataset meets our needs, we can check to see if the data represent the territory required at an appropriate date include the necessary 40 Fundamentals of Spatial Data Quality objects and attributes, but also, if the data have sufficient spatial accuracy or completeness, etc. If that uncertainty is ignored there may be anything from slightly incorrect predictions or advice, to analyses that are completely logical, but fatally flawed. Chapter 3 Approaches to Uncertainty in Spatial Data 3.
Therefore, different types of errors can appear during the data production process see section 2. Quality Components, Standards and Metadata, Sylvie Servigne Nicolas Lesage and Thérèse Libourel. Problems regarding data quality affect all fields that use geographic data. Quality reference bases 3 10. Localization control and global models; 4.
For example, an environmental engineer may need to use a digital elevation model to create a model of a watershed, a land surveyor may need to combine various data to obtain an accurate measurement of a given location, or someone may simply need to surf the Web to find an address from an online map site. The database user must now bear as much responsibility as the originator because the user invokes many of the processing decisions and imparts much of the meaning. This chapter attempts to present certain general concepts related to spatial data quality. Applicable general civil liability regimes. If one polygon does not close properly, it can spread its color all over the map. Chapter 6 presents statistical methods used to measure and manipulate uncertainties related to data.
Detailed description of quality criteria 8 10. Measures on modeling 37 11. Denotation as a radical quality aspect of geographical data. These stages can be prone to problems of definition, misunderstanding, doubt, and error all of which can contribute in various ways to the uncertainty in the information. Data quality is often determined at the moment of data acquisition for example, depending on the technique used. Chapter 1 Development in the Treatment of Spatial Data Quality 1.
Validating the spatial integrity of geographic data. Connectivity in vector data 24 8. Topology and logical consistency; 1. The creation of a Tower of Babel in the form of so many competing standards is perhaps evidence in support of the entropic post-apocalyptic message. In spite of many maps and many models produced before and after August 2005, the inhabitants of the city of New Orleans are now totally convinced that the map is not their territory! In the field of cartography, the term has become important, but the use of this overarching concept is fairly recent. Conclusions and future directions 32 8.
Every map or database is therefore a model, produced for a certain purpose, in which certain elements deemed non-essential have been simplified, grouped, or eliminated, in order to make the representation more understandable and thereby encourage the process of information communication. Quality and metadata as seen by standards 19 10. Completeness concerns the exhaustiveness of a collection of objects. In either case, future trust in the work of the system or the operator can be undermined. Topology and logical consistency 1. This is not to say that there is no concept of reality and truth, just that each term has to be carefully defined and contained inside a system of meaning.
This book and its coverage of the important issues should keep us moving in the right direction. Spatial Data Quality Assessment and Documentation, Jean-François Hangouët. The purpose of the usefulness of quality is then measured by its ability to reduce the uncertainty of a decision. Quality in its context 15 11. Quality and Uncertainty: Introduction to the Problem Chapter 1. Stages allowing external quality evaluation 15 13.
The specifications are a set of rules and requirements that define the way of passing from the real world to the data. Drawing together chapters written by authors who are specialists in their particular field, it provides both the data producer and the data user perspectives on how to evaluate the quality of vector or raster data which are both produced and used. Reasoning Methods for Handling Uncertain Information in Land Cover Mapping. Native projection of a map 27 4. Well-defined objects: error, probability, and Bayes.