1999 Fiscal Year Final Research Report Summary
Spatio-Temporal Data Model
| Project/Area Number |
10680333
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
計算機科学
|
|---|
| Research Institution | Gunma University |
|---|
Principal Investigator |
KANAMORI Yoshinari GUNMA UNIVERSITY, FACULTY OF ENGINEERING, PROFESSOR, 工学部, 教授 (20006266)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
ARITSUGI Masayoshi GUNMA UNIVERSITY, FACULTY OF ENGINEERING, ASSOCIATE PROFESSOR, 工学部, 助教授 (40282412)
|
|---|
| Project Period (FY) |
1998 – 1999
|
| Keywords | Spatio-temporal data model / image sequences / interval / MBR / image sequences / content retrieval / time-interval class / object database / multi-dimensional index |
|---|
| Research Abstract |
(1) We proposed a method for representing numerous spatio-temporal concepts appearing in image sequences. The objects are modeled approximately, by use of cuboids consisting of minimum bounding rectangles (MBR) with time intervals. The spatial relations, including topological and directional relations and the distances between objects, are then defined. The definitions allow for uniform representations of the spatio-temporal queries on objects appearing in image sequences using intervals. We also give examples of such queries for contents retrieval, and illustrate how they are performed in our method.
(2) We describe a way of implementing a spatio-temporal conceptual data model which we have proposed as a class library on an object database system. The class library is composed of three classes : temporal, spatial, and spatio-temporal classes. As the temporal part of our class library, we exploit time-interval class based on time-interval conceptual model which we has proposed. We first build a class for spatial concepts, and then integrate the class with the time-interval class.
We also showed some examples of spatio-temporal queries using class library. Moreover, we investigate an indexing structure based on R*-tree for the class library so that any spatio-temporal query can be efficiently processed with it.
|
Research Products
(8 results)
