3-D Data Acquisition by Trinocular Vision and Shading Information
| Project/Area Number |
62550266
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
計算機工学
|
|---|
| Research Institution | Osaka University |
|---|
Principal Investigator |
YACHIDA Masahiko Faculty of Engineering Science, Osaka University, Associate Professor, 基礎工学部, 助教授 (20029531)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
TSUJI Saburo Faculty of Engineering Science, Osaka University, Professor, 基礎工学部, 教授 (60029527)
|
|---|
| Project Period (FY) |
1987 – 1988
|
| Project Status |
Completed (Fiscal Year 1988)
|
| Budget Amount *help |
¥1,700,000 (Direct Cost: ¥1,700,000)
Fiscal Year 1988: ¥500,000 (Direct Cost: ¥500,000)
Fiscal Year 1987: ¥1,200,000 (Direct Cost: ¥1,200,000)
|
|---|
| Keywords | Comruter Vision / image processing and understanding / 3-D data acquisition / binocular vision / 両眼視 / 三眼視 |
|---|
| Research Abstract |
The importance of 3-D vision is widely recognized in robotics field. For recognition of 3-D shape and measurement of 3-D position, it is important that a vision system can measure the 3-D data of dense points in the scene.
One approach is to measure the distance on the basis of triangulation principle from the disparity of two images. This binoclar vision has, however, a difficult problem that is to find correspondence of features between two images. To solve this problem, we have developed a trioncular vision method which utilizes strong geometrical constraints. We have also developed a new trinocular vision method which establish correspondence among edge segments instead of edge points.
Boundary of curved objects can be classified into two classes; an edge boundary where two surfaces intersect and an extremal boundary which is apparent boundary observed from some view point such as side-surface of cylider. 3-D information of the edge boundary can be obtained by the trinocular vision but not of the extremal boundary. On the other hand, 3-D information of the extremal boundary can be obtained from a single image by shape-from-contour method. We have developed a new method to classify these two kinds of boundaries and obtained 3-D information of curved objects.
|
Report
(3 results)
Research Products
(24 results)
