2000 Fiscal Year Final Research Report Summary
Automatic Production of Highly-Precise Digital Elevation Models from Topographic Map Images and its Application
| Project/Area Number |
11680409
|
|---|
| 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) |
1999 – 2000
|
| Keywords | digital elevation model / DEM / contours / interpolation / topographic map / image processing / Voronoi diagram |
|---|
| Research Abstract |
In this project we made many improvements on the pilot system that can automatically produce a highly-precise digital elevation model (DEM) from images of topographic contour map, in order to make it more practicable. First, we applied the system to a lot of real topographic maps and examined the problems that it has. From the evaluation experiments, we found the following problems in automatic restoration of broken contours : (1) it may fail in restoration ; (2) it cannot detect contours with failed restoration ; (3) it cannot change connection of broken contours once they are connected to restore. To these problems, we could considerably decrease the occurrence of error restorations by using the rigid conditions for restoration and suppressing the generation of unreliable neighboring-contours relationship. Second, we found the problems in the traditional interpolation methods producing a DEM from contours given elevation values, and we developed a new interpolation method using monotone-function. The method has the following features : (1) it can produce a highly-precise DEM ; (2) It does not yield unnecessary vibrations that the other methods like the spline interpolation may often do ; (3) It can generate more suitable surfaces of DEM.We also examined the Grimson's method based on the regularization method, because the methods of this type have not been well investigated. Third, we did an improvement to make the system user-friendly by using graphical user interface : for example, we implemented the interactive restoration of broken contours to utilize the excellent function of human pattern recognition.
|
