1997 Fiscal Year Final Research Report Summary
Estimation of Three-Dimensional Solid object from Ambiguous or Contradictory Views.
| Project/Area Number |
08680394
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Nagoya University |
|---|
Principal Investigator |
WATANABE Takashi Nagoya University, School of Informatics and Sciences, Associate Professor, 情報文化学部, 助教授 (40182927)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
YOKOSAWA Hajime Nagoya University, School of Informatics and Sciences, Professor., 情報文化学部, 教授 (00115651)
|
|---|
| Project Period (FY) |
1996 – 1997
|
| Keywords | intelligent interface / recoqnition of three views / CAD / inconsistent views / estimation of solid objcect |
|---|
| Research Abstract |
While geometry rules relations between multiple views and it present a logical way of three-dimensional solid modeling, it fails to determine a unique object in case more than one solids are constructed from a set of views or the views include inconsistencies. In this research, the prototype system is enhansed to recognize solids with curved surfaces, human processes in the understand-ing of drawing are analyzed, and the falicity to estimate intended objects is improved.
The curved surfaces processed by the current system are cylindrical surfaces, which usual recognition system understand, sperical surfaces and conical surfaces. A new solid modeler is proposed. It is a hyblid modeler which uses both boundary representation method and constructive-solid geometry method. The inconsistencies brought by curved surfaces are divided into three groups. Ressolution methods for each group of inconsistency are presented. In particular, the revison algorithm is improved to remove multiple contradictions in a set of views. The system can resolve inconsistencies and obtain objects even when missing segments are arcs.
This research contribute to investigate the human reasoning process to understand spacial relationship. The result of this research is used to construct more useful recognition system of mechanical drawings. It is also expected to develop an intelligent interface of three-dimensional modeler, which suggests mistakes found in drawings.
|
