Study On Constrained Control And Quality Evaluation Of Land Cover Map

Digital cartographic generalization can be used for focused and clear map production, namely the legible and comprehensive visualization of geographic data at a small scale and for a given purpose, which meets the demands of multi-scale data used for land resource management, planning and decision-making. In the past decades, the research on cartographic generalization had made great progresses in the theories, algorithms, tools, knowledge and rules. Now, to guarantee the generalization quality, how to effectively control the generalization process and how to quantitatively evaluate the quality of the generalized data has become hot topics in cartography.This paper reviewed the latest development of digital map generalization in theories and applications, as well as the research advances of the generalization quality control and evaluation. Take the land cover map generalization as an example, this paper proposed the multi-dimensional constraint strategy and quantitative assessment methods, which were applied to the practical work, that is, the1:10,000land cover map generalization of Gucheng County, Hubei province, to guarantee and measure the outcome quality. The main research work and corresponding results as following:1) Multi-dimensional constraint system and constrained control methods for land cover map generalization were proposed. On the basis of analyzing the influence factors and spatial levels related to generalization, the multi-dimensional constraint system was constructed from five aspects, including semantic constraints, metric constraints, topological constraints, structural constraints and procedural constraints. The executive procedure of constrained control was designed for land cover map generalization, which satisfied the requirement related to control, guidance and examination during the generalization process.2) The framework used for generalized land cover map evaluation was constructed. Four indexes, including map capacity, semantic similarity, structure similarity and pattern stability, and corresponding calculation methods were designed. Through the overall and partial comparison between the original map and the generalized one, the values of index were calculated to measure the quality of the results and the effect of constrained control.3)1:10,000land cover map generalization of Gucheng County, Hubei province, was conducted to get a1:50,000one. The whole area was divided into three parts, that is, the plain area, the hilly area and the mountainous area. Then, based on the method of multi-dimensional constraint system, eight subtasks were performed on the ArcGIS software platform. The results of quality evaluation showed that the map capacity reduces65.14%, the semantic similarity is0.94, the structural similarity is0.91and the pattern stability is0.88. Generally speaking, the effect of generalization control and the outcome quality were completely satisfied, which achieved predetermined goals and cartographic requirement.