Study On Theory And Method For Quality Control Of Digitized Map Data

The quality of GIS data is an important issue related to the reliability of the data and system, and it would influence the benefit of the economy and society on the system. So studying the theory and method for quality control of GIS spatial data is very vital.This dissertation mainly studies some theory and methods for quality control of the scanning digitized map data.First, briefly analyze the error origin about the data of the scanning digitized map. Considering the uneven deformation in scanning map, propose the least square collocation to correct the errors, consequently to eliminate the deformation errors better and improve the quality of the spatial data.Secondly, because the covariance function is so important for collocation, the dissertation uses the minimum of two steps to fit it. Moreover we discuss the covariance function influences on the quality control of the spatial data, which relates to covariance function expression, sample capacity, sampling interval and so on. Through the test, we obtain some meaningful results, which could provide some references to covariance function fitting.Thirdly, because the weights'ratio between observations and signals may be inconsistent, this dissertation also studies the collocation based on the variance component estimation. After analyzing several variance component estimations, we find that this method can adjust the weights'ratio better, and reduce the bad influence caused by the inaccurate apriority information in scanning digitized map.At last, considering the gross errors contaminating the data of scanning digitized map possibly, common collocation can't resist them, thereby we discuss the robust collocation. Especially analyze the covariance function fitting with the robust theory. The test results demonstrate that robust collocation can effectively resist the bad influence of gross errors in digitized map, and make sure the effect of correcting errors is well.