The Comparative Analysis of Three Linear Vector Data Compression Method

Page 1

www.as-se.org/ssms

Studies in Surveying and Mapping Science (SSMS) Volume 3, 2015

The Comparative Analysis of Three Linear Vector Data Compression Method Liu Mao-hua*, Wang Yan, Liu Fang School of Transportation Engineering, Shenyang Jianzhu University, Shenyang, China liumaohua1115@126.com; b31804629@qq.com; chljsunxb@126.com

a

Abstract Aiming at data compression, linear vector data was to be the research object, this paper compared the compression efficiency of three commonly methods: the vertical limit method, interval sampling and Douglas-Peucker algorithm. Fold line with arbitrary contains 17 nodes for the experimental data, using the method of VB bottom of the development of the program design with three kinds of methods, and then the experimental data compression. Program design from the linear data node point, the conclusion of experiments which the program designs can be used as the theoretical basis of data update. Keywords Vector Data; Data Compression; DP Algorithm

Introduction Vector data structure expresses the point, line, surface and other geographical entity accurately with coordinate, it has compact structure, low redundancy, with spatial topological relation, which is convenient for deeper analysis. It is easy to define and manipulate individual space entity, to facilitate network analysis; the output quality of vector data is good, its precision is high and conducive to browse, edit, output spatial data. But vector data redundancy is large, it requires more storage space than raster data. It is also more complex and difficult in the process of editing and processing, so it is necessary to compress vector data. Vector data compression is a basic issue in geographic information system, computer graphics and computer automatic cartography and other disciplines, its essence is a kind of information compression problem. It extracts a bit sequence sets from composition curve point ordered set A, as a new information source for the collection. Within a prescribed degree of accuracy, the collection can express the information of A in the original set in detail. It also can reduce the amount of other unnecessary information in the space, so as to facilitate the preservation of information, saving storage space. Vector data can be divided into graphic elements, punctate graphic elements, and linear planar graphic elements. But from the perspective of compression, the compression of vector data is mainly the compression of linear graphic elements. Because the elements of graphic dot can be regarded as a special linear graphic element, basic graphic elements surface shape is also linear graphic elements, which is surrounded by one or more linear graphic elements. Therefore, the linear graphic elements of the vector data compression become the most basic, the most important issues in the compression Three Common Methods of Vector Data Compression Compression of vector data includes two aspects: one is on the premise of not disturbing the topological relations, for pumping dilute reasonable on the data sampling points; the other is to re-encode the two coordinate data, in order to reduce the required storage space. In this paper, the main research includes commonly used vector data compression methods: vertical limit method, interval sampling, Douglas Peucker vertical limit. Vertical Limit Value Method Vertical limit value method is the calculation of a point away from the adjacent two points of a straight line distance. If it is greater than a certain threshold, this point shall be preserved, if less than a certain threshold, then

10


Studies in Surveying and Mapping Science (SSMS) Volume 3, 2015

www.as-se.org/ssms

that point can be abandoned. The process is applied in addition to the endpoint outside all point of compression, the curve after compression, is obtained. Interval Sampling

Th

re

sh

ol

d

Interval sampling is relatively simple, in the compressed curve, every certain distance, take a reserved in every N point, or a critical value is less than the critical value, the point of all compression, but the first and last point cannot be compressed, certain retention, as shown in Figure 2.

a

a’

b

b’

FIG. 1 INTERVAL SAMPLING

Douglas Peucker Vertical Limit. The basic idea of Douglas Peucker method: first, determine the left and the right of the curve as a starting point (V1 and V6), find out the curve between two endpoints, discrete points from a maximum distance of the two endpoints online point, if the distance is greater than the given threshold, the point is determined to retain. Then respectively with has been the adjacent to keep starting point as the endpoint, use the same method for detecting a discrete point on the curve between them, to determine the next batch of compressed reserve point. Repeatedly used this method until the maximum distance between the discrete points on the curve between the two endpoints much less than the given threshold [9]. Finally meet demand curve can be obtained after compression keeping all points, as shown in figure 2. V3

V3

V4

V2

V4

V2 V5

V1

V6

V1

(1)

(2) Threshold

V3

V5

V6

V4

V3

V4

V2 V1

V6

V1 (4)

V6

FIG. 2 THE PROCESS OF DOUGLAS PEUCKER VERTICAL LIMIT

The Realization of the Three Kinds of Vector Data Compression Method Using VB to develop the underlying vector data compression method, the experimental data on the right side is the line contains 17 points, using three methods of compression, compression accuracy of 1000 twips; get a different compression results, as shown in figure 3. The first is Douglas Peucker compression results, the second is interval sampling compression results, and the last is vertical limit compression results. 11


www.as-se.org/ssms

Studies in Surveying and Mapping Science (SSMS) Volume 3, 2015

FIG. 3 THREE COMPRESSION RESULTS

Part program code of Douglas Peucker are as follows, … For j = first + 1 To last - 1 ReDim Preserve num(j) d = ((py(last) - py(first)) * px(j) + (px(first) - px(last)) * py(j) + py(first) * px(last) - px(first) * py(last)) / Sqr((py(last) - py(first)) ^ 2 + (px(first) - px(last)) ^ 2) num(j) = d Next j For j = first + 1 To last - 1 max = first + 1 If num(max) < num(j) Then max = j Next j For j = first + 2 To last - 1 If num(j) < str Then px(j) = px(j - 1): py(j) = py(j - 1) End If … Analysis of Experimental Results It is not hard to find through the analysis of three commonly used compression methods results: To compress the 17 original data points, the vertical limit method compression casting out the least point which is six and retain enough precision, but there are some points should be abandoned. So the vertical limit method compression can better retain data and compression ratio is not high; Due to the reason for the choice of compression precision, Douglas method after compression curve is three points, the compression ratio is big, that is to say, Douglas method is a relatively ideal compression method. Interval sampling method, it can better save some points, but unable to distinguish between points to the importance of reserved. Conclusion (1) Program design point sampling method, the vertical distance limit value method, Douglas - poke method and so on three kinds of commonly used compression process of vector data compression method. (2) Through the compression of the experimental data, comparison of the Dow Glass Peucker compression is relatively large, can meet the different needs of compression threshold. (3) Experimental base node for processing, the conclusion can be used as the theoretical basis for large amounts of data update, and convenient for application. REFERENCES

[1]

12

Wu Yongying, Cheng Dongliang. “A novel method for implementing multi scale in one database”. J. Huazhong Univ. of Sci.


Studies in Surveying and Mapping Science (SSMS) Volume 3, 2015

www.as-se.org/ssms

& Tech.(Nature Science Edition) 2005,33(7):47-49 [2]

LI Shuai, FANG Yuan-min, XI Wen-fei. “Rapid Compression Algorithm Based on the No Topological Vector Curve”. Science Technology and Engineering. 2011,11(18):4324- 4327

[3] LI Leiding, MA Tiehua, YOU Wenbin. “Analysis of common loss less compression algorithm”. Electronic Design Engineering, 2009,17(1):50-51

[4] HUANG Xingyuan, MA Jinsong. “Geographical Information System Introduction (third edition)”. Beijing: Higher education press,2008,4:48-58 [5]

ZHANG Hong, WEN Lining, LIU Aili. “Basic algorithm of GIS”. Beijing: Science press,2006:21-22

[6]

YANG Dezhi, WANG Jiechen, LV Guonian. “Study of Realization Method and Improvement of Douglas-peucher Algorithm of Vector Data Compressing”. Bulletin of Surveying and Mapping, 2002,7:94-95

Liu Mao-hua: Born in Tieling Liaoning of China, 1981. The major is GIS, master, lecturer. Wang Yan: Born in Qingyuan Liaoning of China, 1979. The major is surveying and mapping, master, lecturer. Liu Fang: Born in Qigihar Heilongjing of China, 1990. The major is surveying and mapping, Postgraduate students.

13


Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.