GISdevelopment.net ---> AARS ---> ACRS 1990 ---> Digital Image Processing

Geo-query: A Geo-object oriented query system integrated with Oracle for GIS

Zhong Zhenxiang, Lu Zhonghui
Dept. of Computer Science,
Peking University, Beijing, China

Zhou Xintie
National Remote Sensing Centre of China


Abstract
This thesis discusses the design of the spatial object-oriented query interface integrated with the relational database management system ORACLE in a Geo-Information System (GIS) Geo-union. The integrated query interface not only relates to the spatial relations between the objects, but also relates to the textual attributes by which the objects are characterized in a relations database. The objects in spatial queries can usually be classified into three categories: point, lien and polygon. Based on this, six basic spatial relations among point, lien and polygon are analysed, and the design of a Geo-object oriented Query Languages (GQL) and a graphical interactive man-machine query interface are presented. The GQL uses an object identifier to connect the spatial information of a object with the textual attributes of the object stored in RDBMS (ORACLE). By adding the map overlay and clip operations in queries, the capability of the query interface is greatly reinforced.

Introduction
The GIS systems the information based on spatial objects (usually divided into points, lien and polygons) in varies kinds of subject maps. In a subject map, each spatial object has different kinds of information: the object's geometric structure and topological relations with other objects (e.g. a series of connected line-segments from a polygon). In addition, textual attributes, for example, an administration district name, the total population of district and the principal governor's name etc. can be attached to polygonal object. These kinds of information are represented in different forms in a computer. In a view of the GQL query system, queries often involve all of these kinds of information. Following example is a typical query:

"find the province names and the names of province governors, where their average annual incomes per person is greater the $300, and the Yangtze river passes through them."

In the sequel, firstly, six categories of spatial relations are discussed, nest the design of GQL is introduced and then the connection between GQL and RDBMS is described briefly, and finally the conclusion is given.

Six categories of spatial relations
  1. Relations between Point-object to Point-Object

    1. Nearest - Point Relation : Find the nearest point-object to the given point in a given map

    2. Farest-point Relation : Find the farest point-object to the given point in a given map.

    3. Point-buffer Relation : Given a point and a buffer size, find all point-objects, which lie within the buffer of the point in a given map.

  2. Relations between Point-objects to Line-Objects

    1. Common Point Line Relations : Given a Point, find all line-objects, which pass thru the same point.

    2. Nearest-to-Point Line Relation : Given a line, find all point-objects, which are lie on the line, in a given map.

    3. Lien-Pass-through Point Relation : Given a line, find all point-objects, which are lie on the line, in a given map.

    4. Nearest-to-Lien Point Relation: Give a line, find the point-object which is nearest to this lien I a given map.

    5. Side point Relation : Given a lien and a side (left or right), find all point-objects which are lie in the same side in a given map.

  3. Relations Between Point Objects to Polygon-objects

    1. Contain Point Polygon Relation : Given a Point, find all polygon-objects, which contain the point, in a given map.

    2. Common - point Polygon objects : given a point, find all polygon-objects, which have this point on their boundaries, in a given map.

    3. Inner point relation : Given a polygon, find all point-objects, which lie within the this polygon, in a given map.

    4. Bound point relation : Given a polygon, find all point-objects, which are on the boundary of the polygon, in a given map.

    5. Outside Point relation : Given a polygon, find all point-objects, which lie outside of the polygon, in a given map.

  4. Relations Between Line-objects to Line-objects

    1. Interest Line Relation : Given a line, find all line-objects, find all polygon objects, which are intersected by the line, in a given map.

    2. Converge Polygon Relation ; Given a line, find all polygon-objects, which contain the start node or end node of the line lies within, in a given map.

    3. Polygon - contain Line Relation : given a polygon, find all line-objects, the whole bodies of which lie within this polygon, in a given map.

    4. Pass- through - polygon Lien Relation : Given a polygon, find all line-objects, which pass through the polygon, in a given map.

    5. Convergence Line Relation : Given a polygon, find all line-objects, whose start node or end node lies within the polygon, in a given map.

  5. Relations Between polygon-objects to Polygon-objects

    1. Island Relation: Given a polygon, find al polygon-objects, which lie within this polygons an Island, in a given map.

    2. Contain Polygon Relation: Given a polygon, find all polygon-objects, which are contained as Islands within the polygon, in a given map.

    3. Polygon - interest Relation: Given a polygon, find all polygon - objects, which overlaps in part with the polygon, in a given map.

    4. Polygon Separate Relation function : Given a polygon, find all polygon-objects which do not overlap with the polygon, in a given map.

    5. Polygon Neighboring Relation : Given a polygon, find all polygon-objects, which adjoin with the polygon, in a given map.
The design Geo-query system
  1. The Eight Function Modules

    In Geo Query System, there are eight function modules i.e.

    • Master Module: Control the system operation and some utilization procedures.
    • Menu Control Module: Menu selection and the menu pop-up and push down.
    • Subject-Attributes Query Module: The query about the subject attributes. Database Attributes Query module: The query about the database attributes.
    • Spatial Relation Query Module: to respond the query about the all spatial relations.
    • Clipping module: To execute the clipping operation and relevant attributes inheritance.
    • Overlaying Module: To execute the overlaying of two maps, and relevant attributes inheritance.
    • Result display module: To produce the query result and the result's display.

    Note that the subject attribute is defined as the principal attribute of a subject map.

  2. The Query Man Machine Interface

    The system adopts the style of menu-driving, integrating pop-up menus of users. The user's Query Screen is divided into six windows as the fig.1 shows.

    Fig. 1 Screen arrangement in Geo-Query

    Notes :
    WIN1: A pop-up window for selecting data values of the subject attributes. All data values of the subject attributes of the current map are listed. User can use the cursor to select many of them.
    WIN2: Selecting database attributes window : A poop-up window. All the database attribute names are listed in the window.
    WIN3: A pop-up window for selecting spatial relations. All spatial relations will be shown in the window according to the goal map's feature and reference map's feature. Note that the goal map is one from which the spatial objects are selected for query results. The reference map is one from which the spatial objects can be indicated and the spatial relations between the goal map and the reference map can be referenced.
    WIN4: Map window. Both goal map and the reference map are displayed in different colors.
    WIN5: Prompt window. Error messages and prompt messages in the process of the query will be displayed. When a query is completed, the query results will be shown on the window.
    WIN6: Operation specification, keyboard usage message and query condition are displayed here. A typical query example is expressed in the sequel.
    Example: find the province names and the names of provinces governor, the provinces should satisfy following conditions :
    1: it's average annual incomes per person (AAIPP)> = $300
    2: the Yangtze river passes through it.
    Analysis:
    The goal map: An administrator map. It's database attributes must include province name, province governor's name and AAIPP etc.
    The reference map: The river distribution map.
    The spatial relation: Line-pass through Polygon Relation.
    Query Steps:
    Step 1: Input the goal map's name thru WIN5, See fig. 2
    Step 2: In WIN 2, using the cursor to select the attribute AAIPP, hen input the condition'>=800, see fig.3
    Step 3: In WIN 3, selecting the term 'Given Ref. Map' and then input the reference map name in WIN5, see fig. 4
    Step 4: Selecting the spatial relation 'Line-Pass-Through Polygon' thru WIN, then using a cross to point out the Yangtze river in WIN4, see fig 5 & Fig 6.
    Step 5: Indicating query condition is completed.
    Step 6: Executing the query, and then display the results in WIN 5, See fig 7.

Fig. 2 Input the map name 'tlanduse in WIN5 as the object map


Fig. 3 A select condition 'ppm>300' of database attr. 'ppm' is given thry. WIN2


Fig. 4 The map 'troads' being given to select as a reference map


Fig. 5 A spatial relation 'Through Fun' means 'Line-Pass-Thru. Polygun' has been selected in WIN3


Fig. 6 A line in color of pink has been selected as a referencing object (a river) of the 'passing thru.' spatial relation


Fig. 7 The results of the query are shown in WIN4 (graphic part) and WIN5 (text part)

The connection of GQL with Oracle Rdbms
In Geo-Query, each object has a unique object identifier given by the system Geo-union. The object identifier plays the role as a unique key field in it's relevant table in RDBMS. By doing so, each object can its all spatial information in geometrical structure files, and in the same way, find its all attribute information in RDBMS through the key field. The correspondence between the object identifier and the key field in relational tables is also automatically maintained by Geo-Query.

Conclusion
The GQL interface for GIS is implemented on the basis of Geo-union system. The running performance and the response time on PC 386 is satisfactory. The system is currently used for a national project of the management information system for forest resources.

References
    Xu Guanhua, Xu Jiyan: The Research Collected works in Regenerating Resource's Remote Senses, Science Press 1988. pp262-279
  • N.S. Chang and K.S Fu: A Relational Database System for Image. "Pictorial Information Systems", Edited by S.K. Chang and K.S, Fu, Springer-Verlag Berlin Heidelberg New York, 1980, pp- 288-321.
  • Albert L. Zobrist and Nevin A. Bryant: Designing An Image Based Information systems. "Pictorial Information Systems", Edited by S.K. Chang and K.S Fu, Springer-Verlag, Berlin Heidelberg New York, 1980, pp 177-197.
  • Y. Edmund Lien and Susan K. Harrig: Structured Implementation of an Image Query Language. "Pictorial Information systems", Edited by S.K. Chang and K.S Fu, Springer Verlag Berlin Heidelberg New York, 1980 pp416-430.
  • Bharat Bharagava: Design of intelligent Query Langauge. "Pictorial Information systems", Edited by S.K. Chang and K.S. Fu, Springer-Verlag Berlin Heidelberg New York, 1980 pp431-445.
  • Michael Stonebraker: Adding Semantic Knowledge to a Relational database System. "On conceptual Modeling", Edited By Michael L. Brodie John Mylopoulos and Joachim W. Schmidt, Springer-Verlag New York Berlin Heidelberg Tokyo, pp333-356.