1998 Fiscal Year Final Research Report Summary
Development of Automatic System Which Produces Highly Precise Digital Elevation Models from Topographic Map Images
| Project/Area Number |
09680387
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
情報システム学(含情報図書館学)
|
|---|
| Research Institution | Iwate University |
|---|
Principal Investigator |
WATANABE Takashi Iwate University, Faculty of Engineering, Professor, 工学部, 教授 (50133905)
|
|---|
| Project Period (FY) |
1997 – 1998
|
| Keywords | digital elevation model / DEM / restoration of contour line / interpolation / topographic map / image processing / Voronoi diagram |
|---|
| Research Abstract |
The main research result of this project is that we succeeded in the development of a pilot system which can automatically produces highly precise digital elevation models from topographic map images. Concretely, we implemented the following computer processings : (1) input of topographic map images ; (2) extraction, labeling and thinning of contour lines ; (3) removal of noise components ; (4) restoration of broken contours lines ; (5) giving elevation values automatically to contour lines ; and (6) production of a digital elevation model from contour lines by using interpolation.
Among the above processings, we made considerable efforts for the two of (4) and (6). For the processing (4), we newly developed a global restoration technique. First, it produces a relationship graph of adjacent contours by making an extended Voronoi diagram. Second, based on an idea of reference contours, it determines pairs of broken contours to be connected and interpolates their gaps with spline curves. In the experiment, we found that it can satisfactorily restore broken contours. For the processing (6), we carried out the detailed validation experiments to investigate accuracy of various interpolation methods by using artificial and real DEM's. The tested methods are linear interpolation, the Akimma's method, the one-dimensional cubic splines, the two-dimensional cubic splines, the FFT interpolation, the surface fitting of cubic polynomial and the morphological interpolation. From the experiments, we concluded that the morphological interpolation and the Akimma's method are excellent, because they have few drawbacks on accuracy and processing time. At present, we can produce precise DEM's easily from topographic map images through a personal computer.
|
