Map Digitized Data Quality Control Theory And Methods

As a decision-making system of spatial relation, the fundamental task of GIS is to analyze and manage spatial date and derivates and abstracts spatial information. The quality of data has decisive function in analyzing and decision-making. So accuracy analysis and quality control of data in GIS is of scientific and practical importance.This dissertation presents some quality control theory and methods on the data of digitalized map. First, the influence property of different kinds of errors in digitalized data and the methods of controlling them are discussed. Some problems on the present methods are also studied. Then the problems of data quality controlling are analyzed from two aspects: statistical method and geometrical method. Especial attention is paid to the project of outlier's detecting, eliminating and controlling in the process of data gathering.According to the quality of digitalized data, the precision assessing model of digitalized map is established. We can test the data's quality from both inner precision and exterior reliability. Both of the accuracy evaluation methods play important role for precision assessing of digitalized data.The existent problem in traditional grosserrors detecting is analyzed. The variance of unit weight is evaluated based on median principle, by which, the influence of unreliable residual is reduced and the reliability of detecting is also improved.Analyzing and controlling the influences of the digitalizing errors are key problems that have to be solved. A functional model (polynomial) is employed to fit the systematic errors and the robust estimation is applied to control the effects of outliers. A mixed method, using the median principle to get the preliminary model parameters and then using the IGGIII scheme for iterative calculation, is proposed in order to ensure the robustness and efficiency of the model parameter estimates.The quality of a digitalized map is usually analyzed or expressed by statistical method, that is, by the standard deviations, variance-covariance or root mean square errors etc. A quite different way to represent the quality of the digitalized map by using the strain tensor is introduced in this paper. It has obvious geometrical meaning.The strain tensor represents the quality of the digitalized map by the variant deformation quantities, that is, local translations, dilations, shears and rotations. All the deformation quantities can be represented in figures shown in the map. The straintensor in representing the quality of a map thus has the advantages of clear geometrical meaning in analysis, obvious visual in representation and simple in calculation. The quality of the map can be analyzed and drown in figures point and point.The least square estimator of the strain tinsor parameters is given. The geometrical meaning of the parameters is explaired and analyzed in theory. It is demonstrated by an example that the strain tensor representation for the quality of the digitalized map is effective.Finally, a deformation correcting system for digitalized map is designed based on MAPGIS. The mixed method that combines systeri errors fit with robust estimation with high breakdown point is used to correct deformed figures and it's validity and feasibility are also be testified.