Research Of Digital Generalization And Key Technologies In Production Processes Based On New DEM

Digital Elevation Model (DEM) is one kind of the national products of geography information, which has been widely applied in many departments about the earth’s three-dimensional information such as water resources, mapping, national defense and others, and is the most important part of DGDF. In half of a century’s development, it already has a very large-scale production in practice and has played a significant role in many applications.For many-years’efforts, the National Survey and Mapping Department has build the DEM databases in scale of 1:10000,1:50000,1:25000. The follow work is how to keep the quality and update the databases in time, which is the focus for many users and the mapping workers.DEM, as a digital representation of terrain surface, has many advantages that the traditional topographic map can not compare with. But it also has much problem in practice, especially in quality assessment. So we should take the theory basis and the technical basis to think through their shortages. As we know, from the definition to the production, from the real terrain to the quality assessment, every process is closely related. So in this situation, the new DEM theory and technology in application have given us a new way to solve all these problems.At present there are many kinds of data sources and application requirements, then there presents many discussions about data organization in multi-scale and sampling of elevation. So in this paper, we do some researches about the processive digital generalization and its technology based on three characteristics of terrain, at the same time, we also do some data experiments on large-scale DEM production generation. The material works as followed:Chapter One: We have analyzed researches at home and aborad about the traditional map generalization/DEM generalization, how to reserve terrain features, DEM quality standards, DEM quality assessment, the visualization in multi-scale, the production process and so on. For there are so much problem, we find the new DEM is another effective way for DEM generation. There are two basic principles:(1) the high fidelity DEM is the base and the guarantee for the DEM generalization; (2) To generate the high quality DEM in appointed resolution, it should be through DEM digital generalization. Especially in the feature-reservation problem, we take the grid method based on map algebra through the bottom to the top, which is different from the traditional----digital generalization in feature modeling. We compare with the adjacent elevation to judge which is feature point or in the feature tree. We find it is a effective new way.Chapter Two:There are three characteristics of terrain:(1) the accuracy of elevation; (2) the right order of elevation; (3) the feature structure of elevation. In the practice and DEM development, three characteristics are independent and interrelated, we should understand the relationships among them. They are the base and the prerequisite for the digital generalization.Chapter Three: In feature modeling, we find the "feature reservation" is the key to the problem, so based on this point, we take the levels and system structures of features to consideration, then find the relationship of terrain features, and reserve these important features in different levels. And finally we can do the DEM generation through digital generalization easily. Some data experiment results show that for different types of terrain, this algorithm is nice and practical.Chapter Four: As we know, the problem of contour lines generalization is a century challenge. If we want to solve this problem, we should consider the structure in three dimension. Since the "digital generalization" is based on the three dimension model. Such DEM is equivalent to the topographic map in resolution at 0.1mm. So we can take two results from one "digital generalization" theory: one is DEM generalization, the other is contour lines generalization. At the same time, we do some data experiments, and we also find the results are very ideal.Chapter Five:Based on the previous results and the technology of "digital generalization", we can integrate those softwares to form a new process for high-quality DEM generation. In this paper, we have discussed about the technical processes, specifications and the two core technologies. And we have done large-scale data experiments, the results show us the new process is feasible and mature in application.Chapter Six:Through the large-scale data experiments results, we have done much anaysis on the quality assessment under the new quality standards. On one hand, we do the compare in multi-scale visualization; on another hand, we make some tools to do quantitative statistics to show the new quality standards through figure. The final statistic results tell us the new theory and technology of DEM is feasible and superior.