COP-VW: Cone-over-Projection Directional Model Viewer

Jefferson Amaral da Silva

and Karla Donato Fook

State University of Maranhão (UEMA)

São Luis-MA, Brazil

Department of Informatics, Federal Institute of Education, Science and Technology of Maranhão (IFMA),

São Luís-MA, Brazil

Keywords: Directional Relationships, Geometric Algorithms, Geographic Information Systems, COP-Model, COP-

VW.

Abstract: An important feature for geographic systems is to identify directional relationships between objects stored

in Spatial Database Management Systems (SDMS). This feature is used with other information type, such as

Metric and Topological Relationships. This features group provides relative positioning between

Geographical Objects in some direction, and supports decision makers job, such as to decide where setting

up certain buildings, opening of access roads, and installation of transmission towers, for instance. Some

approaches addresses to direction definition, but without implementation. Cone-Over-Projection Directional

Model (COP Model) refines directions definition and enhances spatial queries accuracy in a significant

form. The Model has a computational tool to support its application. This work presents the COP Model

implementation through COP Viewer (COP-VW) application that uses Geometric Algorithms added to an

SDMS located in a Web Server. Several users can access COP Model resources at different geographical

locations, and different geodatabases can be available to environmental and urban planning decisions. This

is a first step to migrate COP Model to mobile platform.

1 INTRODUCTION

The identification of Directional Relationships

between objects stored in Spatial Database

Management Systems (SDMS) can set a useful

source of information. This feature combined with

other, such as Metric and Topological Relationships

supports urban planning job. Decision-makers use

this features to decide where to make buildings, to

open access roads, and to install transmission

towers, for example. In these cases, is fundamental

to know the relative positioning between

Geographical Objects in a given direction.

Depending on the application domain, some

spatial relations may be more significant than

others (Papadias and

Theodoridis, 1997).

Topological relationships have a set of definitions

and operations widely accepted (Egenhofer et al.,

1991; Egenhofer and Franzosa, 1995). There’s no

unified Directional Relationships concept. This point

is a problem and causes different approaches to

define directions, each with its positive and negative

points (Theodoridis et al., 1996).

As result, several models define the directional

relationships, each one with their own

characteristics. Researchers present Models in

different implementation. Silva and Fook (2013)

propose Cone-Over-Projection Directional Model

(COP Model) to refine direction definitions. COP

Model is a hybrid model that includes characteristics

of the models based on Cone (Peuquet and Zhan,

1987) and Projection (Frank, 1992).

We highlight that the COP Model enhances

spatial queries accuracy in a significant form, and

aims to support decision makers in their resource

management activities. This paper presents COP-

VW, the computational tool that supports COP

Model use. This software uses a Geodatabase hosted

in a Web Server. This Geodatabase allows free

access to information which supports decisions

concerned to environmental and urban planning. In

this case, makers’ decisions can visualize several

databases in different places around the world. There

is a proposal to upgrade this Geodatabase to Mobile

platform.

Next section presents directional relationships

between objects approaches, and the COP Model

245

Amaral da Silva J. and Donato Fook K..

COP-VW: Cone-over-Projection Directional Model Viewer.

DOI: 10.5220/0005449602450252

In Proceedings of the 11th International Conference on Web Information Systems and Technologies (WEBIST-2015), pages 245-252

ISBN: 978-989-758-106-9

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

adds. Section 3 presents COP Model and their

characteristics. Section 4 presents COP Viewer

specification, application, results and discuss about

COP Model use. Finally, Section 5 presents some

conclusions about the geographical application.

2 RELATED WORK

Authors propose directional relationships between

objects approaches. There are two categories which

include basic models for defining directional

relationships: Cone-Based Models and Projection-

Based Models (Xia et al., 2007).

The Cone-Based Models partition the space by

using lines with an origin angle α (Figure 1). Typical

models include the 4-direction Model, Figure 1(a),

the 8-direction model, Figure 1(b), and the triangle

model, Figure 1(c) (Tang et al., 2008). These

Models provide an accurate identification of

directional relationships for point geometries.

However, misleading directional relations may be

produced when reference objects are lines or

polygons (Tang et al. 2008).

Figure 1: Cone-Based Model.

The Projection-Based Models divide space by using

lines parallel to the axes (Spiros et al., 2007). The

space around an object reference A is partitioned

into nine areas:

 north (NA)

 northeast (NEA)

 east (EA)

 southeast (SEA)

 south (SA)

 southwest (SWA)

 west (WA)

 northwest (NWA)

These areas refer to the cardinal and ordinal

directions. There is one extra region corresponding

to the Minimum Bounding Rectangle (MBR) of the

reference geometry (0A), as shown in Figure 2. In

this category, the MBR Model is prominent (Tang et

al., 2008).

Figure 2: Projection-Based Model.

At first, Egenhofer et al. (2000) introduced a

Minimum Bounding Rectangles (MBR) Model,

which uses the 4-intersection matrix, projecting a

grid over the concerned geometries (Egenhofer and

Herring 1991). This Model expresses directional

relationships between the MBR of the reference

object A and the primary object B. It considers A

(interior and boundary), and B (interior and

boundary). Later, authors extended the model to 9-

intersection matrix. Now the approach considers

interior, boundary and exterior of A and B objects

(Egenhofer and Herring 1991). Minimum Bounding

Rectangles Model is not suitable for treating points,

since they do not actually have MBR. Besides,

Egenhofer work was theoretical, not presenting

implementation.

Zhu et al. (2012) presents a model for defining

directional relationships between geometries based

on Geo-Ontologies. In this model, secondary queries

made on Geo-Ontologies settle the directional

relationships. The Zhu's model adds semantics to the

research and enables tapping knowledge about the

directionality in the objects represented in the

ontology. However, this addition implies data

arranged in the form of ontological basis on the

studied area, and could result in spatial databases

becoming incompatible with this model, if there is

no ontology regarding their spatial context. Further,

the ontological database is external to the SDMS,

resulting in the need for two databases, one spatial

and one ontological, separated to perform the search.

Thus, integration with the existing SDMS resources

becomes problematic and complex. In this work, the

model is theoretical, and the implementation will be

treated as a future work. Next presents COP-Model,

which aims to improve the accuracy of spatial

queries significantly.

WEBIST2015-11thInternationalConferenceonWebInformationSystemsandTechnologies

246

3 COP-MODEL

Cone-Over-Projection Directional Model is a model

for defining directional relationships between two

geometries. For this, the model uses the union and

adaptation of two models: 8 Directions Cone-Based

Model and MBR Model (Projection-Based)

introduced by Egenhofer et al. (2000).

COP Model simulates the human view angle,

starting from a central point and grows as it moves

away from the source with the ability to treat line

and polygon geometries from Egenhofer model

(2000).

The model developed core was overlapping the

grid projection with the conical grid to identify areas

where they differ. This feature reduces the gap

between sizes of existing partitions on the projection

model. The union of cone and projection models

uses concepts and algorithms from computational

geometry, and processes how objects relate to each

other and with the regions defined by the designed

model. COP Model improves significantly the

accuracy of spatial queries. This feature is essential

to support the territorial planning applications.

Da Slva and Fook (2013) developed a pseudo-

implementation of COP Model. This pseudo

implementation is generic and allows to be

translated into various programming languages and

in third-party application or database. These features

were incorporated to SDMS. Developers can use this

module as standalone software or as an extension of

some existing Geographical Information System

(GIS) as gvSIG, Kosmo GIS or QGIS. This

implementation allows you to check the COP Model

application and is a differential feature over others

approaches.

Next section presents COP-VW, a Geographical

Information System developed as a COP Model

concept proof.

4 COP-VW

COP Viewer or COP-VW is software that allows

COP Model application. Their architecture contains

a Database Server. Several users in different

geographical locations can access Geodatabase

through Web to make spatial queries using COP-

VW (Figure 3). The GIS allows that different

decisions to be made for different territorial

purposes more precisely. For example, select area to

environmental protection or build factories or

hospitals.

Figure 3: COP-VW Architecture.

We used the Java programming language, Sun Java

Windowing (SWING) and Java 2D Graphics API

(Java 2D) as follows:

 Java APIs such as Java Database Connectivity

(JDBC) to connect with SDMS

 SWING to create the GUI

 Java 2D for graphic display of retrieved

information from SDMS

There are several ways to implement new

functionalities at Spatial Queries. We implement

COP Model features as SDMS extensions from their

source code. Geographical Information Systems

only need to invoke operations aggregated to the

database to use COP Model features (Figure 4). We

stress this point as strong feature of work.

Figure 4: COP Model extensions insertion layer (Silva and

Fook, 2013).

Spatial queries can unite topology and distance

concepts with directionality concepts on the same

database. COP-VW uses data structures of

standardized SF-SQL hierarchy by OGC.

COP-VW:Cone-over-ProjectionDirectionalModelViewer

247

Figure 5: COP-VW main Form.

4.1 Specification

COP-VW works object concepts from COP Model.

For this, we created classes to represent the

information brought from SDMS. These classes

were generated internally in the COP-VW. The

model has two classes, and the spatial reasoning

complexity was abstracted. We use standard Java

programming language API and API provided by

PostgreSQL/PostGIS.

The GIS includes resources to handle geospatial

entities, objects and geometries, as well as a

Descriptor Table. COP-VW uses SDMS and COP

Model functionalities. There is not spatial processing

functionality in it.

4.2 Implementation

COP-VW releases user to write SQL queries/SF-

SQL directly. There are Forms to generate spatial

queries visually. Figure 5 shows the main COP-VW

Form. There are several parts in this Form: Figure 5

(a) displays information about the database

connection and projection used by the geometries.

Figure 5(b) displays the spatial query produced by

the tool. In Figure 5(c) there are creation and submit

queries buttons. Figure 5(d) shows spatial layers

presented in the open database. Figure 5(e) has

graphical representation of the loaded geometries

from database and results of spatial queries

submitted. Finally, Figure 5(f) we see textual details

recovered from spatial queries.

The display panel of the geometries shown in

Figure 5(e) provides translational and scale features

of the displayed geometry. This allows a better

evaluation of the consultations results.

User can create Spatial Queries in a COP-VW

Form (Figure 6). There are two different categories

of queries available: “Test Target” and “Find

Target”.

To perform “Test Target” type queries, user must

inform: Reference Object, Target Object, and any

direction of the COP Model. After that, COP Model

implementation calculates the percentage of the

area/length of Object Target in the direction

informed towards the Reference Object.

Figures 6 and 7 show the Form for this query

creation and its result, respectively.

Figure 6: COP-VW “Test Target” query Form.

The COP-VW Query Text produced in Figure 6 is a

SF-SQL: SELECT

B C

WEBIST2015-11thInternationalConferenceonWebInformationSystemsandTechnologies

248

Figure 7: COP-VW result for “Test Target” query.

stx_cop_north_east(E_REFERENCIA_geo

metry, E_ALVO.geometry) AS “value”

FROM bairro E_ALVO, bairro

E_REFERENCIA WHERE

E_REFERENCIA.id_bairro = 50 and

E_ALVO.id_bairro = 115.

Figure 7 shows output generated by query “Test

Target”. This query consists of a visual pointing

Reference Object (magenta), Object Target (green)

and grid lines Projection and cone.

User can make another category query using

COP-VW: “Find Target”. Figure 8 displays the

Form to generate this query type.

Figure 8: COP-VW “Test Target” query Form.

“Find Target(s)” queries need from user: object

reference, a direction in which it will check for

potential targets and the layer in which it will carry

out the search. Application also requests a percent

threshold for an object to be really considered valid

target, because the percentage of area/length of the

targets is considered.

Such consultations even offer the possibility of

adding Topological and/or Metric Relationships

tests. Thus, search results can be refined according

to these characteristics of Spatial Objects. After that,

COP-VW requests the SDMS to find Spatial Objects

with requirements and limitations imposed.

An SF-SQL query example generated by this

COP-VW Form is:

SELECT E_ALVO.texto AS

"nome", E_ALVO.id_bairro AS "id",

stx_cop_north_east(E_REFERENCIA.geometr

y, E_ALVO.geometry) AS "porcentagem"

FROM bairro E_ALVO, bairro E_REFERENCIA

WHERE E_REFERENCIA.id_bairro = 50

AND

stx_cop_north_east(E_REFERENCIA.geometr

y, E_ALVO.geometry) >= 0.3

AND NOT

(st_touches(E_REFERENCIA.geometry,

E_ALVO.geometry))

AND st_distance(E_REFERENCIA.geometry,

E_ALVO.geometry) > 1.0.

Figure 9 presents the result of this query

example.

The COP-VW also allows queries submission

that relate different types and different Geometric

tables, providing consultations involving the

Polygon/Polygon, Polygon/Line, Polygon/Point,

COP-VW:Cone-over-ProjectionDirectionalModelViewer

249

Figure 9: COP-VW result for “Find Target” query example.

Line/Polygon, Line/Line and Line/Point pairs. We

carry out these cases in the SDMS extension and

COP-VW only invoked them.

4.3 Discussion

Projection grid tends to partition the

disproportionately space in that it moves away from

the origin, favoring Side directions. The COP Model

managed to minimize this by superimposing a

conical grid. There were growth areas of the

partitions to verify divisions generated by the COP

Model. For this we used the same parameters used in

the projection grating, or square with 1 unit MBR

side and offset from 1 to 10 units away from the

origin. Figure 10 shows the percentage distribution

of space between the cardinals, side directions and

added by COP.

Figure 10: Growth Chart of directions in COP Model.

We show in Figure 10 an even exponential growth

of the area presenting the Side directions and

straight to the Cardinals directions. However

directions added by COP Model also grow

exponentially, which balances the growth of side

directions, so this is not reaching the 59% of the

partitioned area, as shown in Figure 11.

Figure 11: Percentage of space occupied by the COP

Model directions.

According to user interpretation, areas of the

cardinal directions with their respective COP

directions can be added, and the COP directions are

influenced by the cardinal directions. These

considerations balance cardinals and collaterals.

Figure 12 shows a comparison between the

percentage areas occupied by cardinals and

collaterals directions with the projection grid and the

COP Model grid. We have the values related to

WEBIST2015-11thInternationalConferenceonWebInformationSystemsandTechnologies

250

cardinal directions and their respective influenced in

the last one.

Figure 12 shows that with the COP Model grid,

the growth of side areas is no longer exponential and

Figure 12: Comparison between Projection and COP

Model partitions.

is replaced by closer linearity behaviour. Also be

seen that as it moves away from the source, the gap

in favour of the side direction continues to grow, but

in smaller and smaller quantities. On the whole, the

difference in areas that time had reached 82% in the

Projection grid fell to 16.78% with the grid of the

COP Model in this given situation. Thus, the model

reduced the discrepancy between the areas by a

significant way.

5 CONCLUSIONS

As a result of this work we granted some tangible

artefacts: The mathematical definition of the COP

Model, the COP ME (a suite of products that form

the seamless deployment to SDMS), and COP-VW

Geographical Information System.

The COP-VW allows to verify the COP Model

implementation via Web, and to show that spatial

processing functions carried out in the COP-ME are

usable by a third-party software without needing

redeployment or any specific suitability.

The presented GIS is the COP Model

implementation from visual and friendly way. The

implemented features could be tested individually

and with existing features in SDMS, such as on the

topological and metric relationships, so it was

indeed found the implementation was faithful to the

model definition. Further, the COP-VW can also be

used as a graphic viewer geometries stored in any

database that uses PostgreSQL/Postgis, regardless of

whether or not using the COP-ME extension. COP-

VW works with a Geodatabse located in a Web

platform. This feature allows free access to

information and supports decisions related to

environmental and urban planning. Decision makers

can visualize several databases in different

geographical locations. This point is a strong feature

of work.

As further work, we intend to develop the COP-

VW in a Mobile platform. Finally, we stress that all

procedures performed and created products were

made using technologies free/opensource, which

made the COP-VW free of the need to pay license

fees, both for developers and for those using.

REFERENCES

Egenhofer, M., 2000. Qualitative Spatial-Relation

Reasoning for Design. National Center for Geographic

Information and Analysis, Department of Spatial

Information Science and Engineering Department of

Computer Science. University of Maine Orono, USA.

Egenhofer, M.; Herring, J., 1991. Categorizing Binary

Topological Relationships Between Regions, Lines,

and Points in Geographic Databases. Orono, ME:

Department of Surveying Engineering, University of

Maine.

Egenhofer, M.; Franzosa, R., 1995. On the Equivalence of

Topological Relations. International Journal of

Geographical Information Systems, v. 9, n.2, p. 133-

152.

Frank, A. U., 1992. Qualitative spatial reasoning about

distances and directions in geographic space, J. Visual

Lang. Comput., 3, 343–371.

Papadias, D., Theodoridis, Y., 1997. Spatial Relations,

Minimum Bounding Rectangles, and Spatial Data

Structures. International Journal of Geographic

Information Science 11(2), 111-138.

Peuquet, D., Zhan, C. X., 1987. An Algorithm to

Determine the Directional Relation Between

Arbitrarily-Shaped Polygons in the Plane, Pattern

Recogn., 20, 65–74.

Silva, J. A. da and Fook, K. D., 2013. Addition of the

Directionality Concept in Spatial Queries on SDMSs

Using the Union of the Cone-Based and Projection-

Based Models. Proceedings of XIV Brazilian

Symposium on Geoinformatics - GEOINFO,

November 24-27, Campos do Jordão, Brazil. ISSN:

2179-4820.

Spiros, S., Nikos, S., Timos, S., Manolis, K., 2007. A

Family of Directional Relation Models for Extended

Objects, IEEE Transactions on Knowledge and Data

Engineering, 19(8), pp.1116-1129.

Tang, X., Meng, L., Qin, K., 2008. Study On The

Uncertain Directional Relations Model. Based On

Cloud Model. The International Archives of the

Photogrammetry, Remote Sensing and Spatial

Information Sciences. Vol. XXXVII. Part B2. Beijing.

Theodoridis, Y., Papadias, D., Stefanakis, E., 1996.

Supporting Direction Relations in Spatial Database

Systems, Spatial data handling International

Symposium. In 7th, Spatial data handling, ISBN:

COP-VW:Cone-over-ProjectionDirectionalModelViewer

251

0748405917.

Xia, Y., Zhu, X., Li, D., Qin, K., 2007. Research on spatial

directional relation description model, Science of

Surveying and Mapping, 32(5), pp.94-97.

Zhu, X., Chen, D., Zhou, C., Li, M., Xiao, W., 1012.

Cardinal Direction Relations Query Modeling Based

on Geo-Ontology, State Key Laboratory of

Information Engineering in Surveying, Mapping and

Remote Sensing, Wuhan University, International

Archives of the Photogrammetry, Remote Sensing and

Spatial Information Sciences, Volume XXXIX-B2,

2012 XXII ISPRS Congress, 25 August – 01

September 2012, Melbourne, Australia.

WEBIST2015-11thInternationalConferenceonWebInformationSystemsandTechnologies

252