Quality Evaluation Of Linear Features' Simplification Algorithms

With the research on automatic cartographic generalization developing and extending continually, quality of cartographic generalization is becoming more and more important, if lacking quality control, automatic generalization can't be realized completely. Linear features take eighty percent of cartographic features, and simplification algorithm is regarded as the most important algorithms in map generalization and multi-scale expressing and processing of spatial data in GIS, therefore, it's significant to evaluate linear features' simplification algorithms.On the basis of taking constraint conditions as the strategy of quality evaluation, linear features' simplification algorithm is evaluated from four aspects which are spatial relationship, geometric accuracy, shape characteristic and applied range, and credible conclusions are got by a great deal of experiments, the main work and creations are as the following:1. Evaluation flow of linear features' simplification algorithms is put forward. The flow summarized the problems on geometric and semantic aspects easily appearing in the process of linear features' simplification, and established constraint conditions as the guidance of evaluation, besides, combining the characteristics of linear features, evaluating figures are given initially.2. Consistent evaluating rules of topology about lines' intersection are designed. Idea of designing consistent rules of topology is offered, and evaluating figures of topology intersection check are proposed.3. Geometric accuracy of linear features' simplification algorithms is evaluated. That the influence of linear features' accuracy after simplification can be divided into geometric accuracy and attribute accuracy is analyzed, a series of evaluating figures of geometric accuracy are proposed, several typical linear features' simplification algorithms are evaluated and some credible conclusions are given.4. Error propagation model of linear features' simplification algorithms is developed. Researching method of simplification algorithms' error propagation is proposed, and the model is visualized by error ellipse.5. Shape of linear features is evaluated based on fractal theory. General method of evaluating curves' shape based on fractal theory is introduced. A new method of identifying curve's bend is proposed, algorithm applying the new method can keep the shape of curve better.6. Influence of parameter to simplification algorithm is analyzed. Parameters of different algorithms are given to simplify rivers and roads and performance of parameters is analyzed briefly, applied range which is aiming at constraint conditions are given initially from four aspects that are spatial relationship, geometric accuracy, shape characteristic and applied range.