一种基于先验知识的矢量边界更新方法

A Method Based on Prior Knowledge for Vector Edge Update

本文针对GIS中地物矢量边界更新繁琐、低效的问题,提出一种以GIS数据作为先验知识的简便矢量边界更新方法。该方法针对GIS线性矢量边界只需有限个拐点描述的特性,将对边界寻找转化为直接对矢量边界上拐点位置的确定,简化了变化检测与更新过程。以TM遥感影像为数据源,矢量shape文件为先验知识,分别利用该方法检测、更新扎龙湿地内克钦湖、东升水库以及一块沼泽地边界。并运用简化的缓冲区检测算法计算拐点更新准确率,同时对本文结果与手动更新结果的相似性进行计算与分析。实验结果表明,本方法不但简化了变化检测和更新过程,而且可以得到接近手动更新方法的准确率,并且更新后矢量多边形与手动更新所得到的矢量多边形保持了较高相似性,达到了较好的效果。

Data update is an important task for GIS.However,complex and inefficient process has been a principal problem that gets in the way of updating GIS data in time during vector data update in GIS,it is necessary to substitute an automatic method for manual method currently.In order to make data update less dependent on arbitrary judgement,auxiliary information should be cited.During the change detection of remote sensing images, GIS data have played an important role as prior knowledge in increasing accuracy and simplifying process.Approaches of change detection and data update has been proposed in recent years.However,those approaches are based on neural network,knowledge base instead on linear vector data.Linear vectors that describe the object region edge in GIS,is usually irregular polygon which is represented as connection of the discrete points on the edge.Hence,it makes sense that the method depending on the attributes of linear vectors above is figured out to update the coordinate values of points on the boundary.A simple and convenient method,which utilizes data in GIS as prior knowledge,is proposed here for vector edge update to solve the problem of complexity and inefficiency.This method involves three-step processes of searching edge: generally locating points,recognizing the convexity and concavity,and detailedly describing the polygonal shape.The points moves within the limited regions that threshold values determines.This method integrates remote sensing images and prior GIS data,as well as replace searching edge curve with locating inflexions to simplify change detection and update process.In addition,the method points out an idea of integration of remote sensing and geographical information system data with the carrier of linear vectors.With the experimental data of Landsat TM remote sensing images and the prior knowledge of vector data saved as shapefile of ArcGIS software,the method is applied to detect and update the edges of Keqin Lake,Dongsheng Reservoir and an area of marsh in Zhalong Wetland in Heilongjiang Province,China.And then update accuracies are tested by the simplified buffer detection algorithm for computing polygonal error based on original buffer detection algorithm,and the polygonal comparability between this presented method and the manual method is calculated and analyzed.In terms of the characteristic of polygon,five factors are adopted in evaluating the polygonal comparability: major direction angle and minor direction angle of minimum exterior rectangle,length and width of minimum exterior rectangle,and polygonal perimeter.Relative error of these five factors are computed,based on which mean value is considered as evaluation of polygonal comparable percentage.Besides,different from original buffer detection algorithm,a simplified buffer detection algorithm is pointed out.Based on the improvement,buffer detection algorithm can be calculated through infinite numbers of points on the polygon that stand for object region edge.Estimations are obtained according to those two criteria above.The results show that the update accuracy of proposed method reach above 80%,a little lower than the result derived from manual method,though good polygonal comparability very close to the result of manual method.In spite of that,the approach presented performs an automatic procedure of vector edge update in GIS.Apparently,this method well simplifies change detection and update process,as well as achieves an update accuracy close to the manual method,and a relatively satisfactory comparability with the manual update method.Besides,the results implies that different experimental objects lead to distinct evaluated results.The simpler the shape of edge is,the better the method is evaluated.Thus,the result from experiment on Keqin Lake is better than Dongsheng Reservoir and marsh,as the contour of Keqin Lake is simpler than that of the other two.Nevertheless,there are still several disadvantages to be improved.This method loses control of contour shape of object so that the movement can not be restricted to be harmonious between adjacent points.Also,the moving range threshold of points are supposed to be settled by researcher,which means that the automatic process is affected by human idea.