Methodological Research On Automatically Recognizing And Extracting Morphological Cells Of Linear Feature

Improving automation and intelligent of handing the spatial information is the main trend of Geo-Information technology.Automatically recognizing and extracting linear feature's morphological cells from massive map data is the core content of the trend,which will be significant to map data compression?map automatic generalization,Geo-spatial data matching?analysis and mining.This article taking two typical linear features,river and road for example,study a general method of recognizing and extracting linear feature's morphological cells which exploit the boundary information of linear feature.The main content and achievement of this article as followed:(1)The relative principle?method and measure index that refer to linear feature's morphological cells has been dicussed.Based on the examples of river and road,this article make a conclusion and summarizes the correspondence between the spatial structure feature of linear feature and the data set of digital map which record it;states the indicator system that measure the morphological features of linear feature in different level;based on semantic level,a classification system that used to distinguish the type of linear feature's morphological cells has been built,and which mainly consisted of criterions of automatically recognize the linear feature's morphological cells.In the above process,aimed at the issue of recognize the river's morphological cells,a new identification parameter that distinguish the type of river bend and a new thinking that use the shape and number of sandbank to recognize the furcated river were proposed;and aimed at the issue of recognize the roads' morphological cells,states the principle that use feature points to recognize roads' junctions and curves,and proposed three properties that those feature points should have in a Double-boundary-road-network Map.(2)An algorithm to recognize and extract junctions and curves of roads has been designed.By using the indexes such as the steering angle between two adjacent line segments in one of the two road boundaries,the average steering angle per unit length of road boundaries,this algorithm can overcome the limitation of currently presented related methods and corresponding algorithm in solving the same problem,and which can successfully find the location and judge the type of junctions and curves of roads.The algorithm use the spatial relationship of feature points which picked out by the above indexes,determine the location of a road's junction or curve;use the detail information of a road's junction or curve,judge its type.The algorithm still can recognize and extract the atypical junctions and the complicated junctions by decomposed them into typical junctions.(3)Use a double-boundary-road-network map(scale:1:500)of Fuzhou city(China)as a sample,accomplished the empirical study of the method of recognizing the junctions and curves of roads.The experiment shows that the proposed algorithm is superior to the presently available ones with respect to robustness,efficiency and automation.When there are exist some obvious quality problems,such as data missing,spatial location errors and topologic errors in the road-network map data,and on the condition of recall ratio approach to 100%,the precision ratio is above 80%.In particular,to recognize the main roads,the precision ratio of the algorithm can approach to 100%.(4)Accomplished the application study of using the algorithm to support driving and map automatic generalization.The former study use some detail information of a junction,such as the central point?mid-line,which was extracted by the algorithm,and by calculate the critical speed when a car pass the junction,then make a early warning if the current speed of the car is faster than the critical speed.The latter study taking advantage of the correspondence between the size of the junctions of roads and the road grade,and following the ascending order,implemented the deletion algorithm of deleting junctions,and meanwhile,the algorithm also delete the relative road sections which connect to the deleted junctions,and therefore accomplished one of the core missions that need delete the abandoned roads on the process of map automatic generalization(the algorithm can easily maintain the connectivity of the road-network).