| Project/Area Number |
11680380
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Tokyo University of Mercantile Marine |
|---|
Principal Investigator |
OSHIMA Masaki Tokyo University of Mercantile Marine, Faculty, Mercantile Marin Scie., Professor, 商船学部, 教授 (50242318)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
ZEN Heito Tokyo University of Mercantile Marine, Faculty, Mercantile Marin Scie., Assistant Professor, 商船学部, 助教授 (20216568)
|
|---|
| Project Period (FY) |
1999 – 2001
|
| Project Status |
Completed (Fiscal Year 2001)
|
| Budget Amount *help |
¥3,600,000 (Direct Cost: ¥3,600,000)
Fiscal Year 2001: ¥1,100,000 (Direct Cost: ¥1,100,000)
Fiscal Year 2000: ¥1,100,000 (Direct Cost: ¥1,100,000)
Fiscal Year 1999: ¥1,400,000 (Direct Cost: ¥1,400,000)
|
|---|
| Keywords | Vision / outdoor scene / ship image / optical flow / sonar image / reconstruction / plant image / wave image / outdoor scene / Vision / plant imdge / stereo |
|---|
| Research Abstract |
In this research, to establish basic knowledge to extract and find application of useful information from outdoor scene image are aimed. The target objects are, for example, ships as moving artificial object and outdoor natural objects. The results can be categorized to ones on the sea and ones under sea. On the sea : To extract moving vector from ship image sequences, usually two assumptions are used : equi-brightness on the same point and smoothness. Instead of the later, local assumption of equi-movemerit is used and was successful. Extraction of moving ship from observed image sequence on moving ship has been attained on the inference of ego motion. Experiments to measure the direction of waves are performed, where edge detection of wave is used as basic technique. An approach to extract oil slick from the SAR image observed from satellite is performed. Under the sea : Reconstruction of sea ground structure from multiple side scan sonar images are investigated. After extracting common area within two images (oppositely observed), multi-stage matching to compensate possible distortion associated !by the observation. The result of the reconstruction of sea ground seems satisfactory. Others: Stereo I matching on plant image applying of increment coding matching is successful and some reconstruction result is convincing. In general, our aim is so far attained.
|
