2005 Fiscal Year Final Research Report Summary
Chronological Analysis of Wind Erosion of Kitakami Highland using Satellite Images
| Project/Area Number |
16510001
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Environmental dynamic analysis
|
|---|
| Research Institution | Hirosaki University |
|---|
Principal Investigator |
IIKURA Yoshikazu Hirosaki University, Faculty of Science and Technology, Professor, 理工学部, 教授 (30109897)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
SAITO Hiroyuki Hirosaki University, Faculty of Science and Technology, Associate Professor, 理工学部, 助教授 (70264091)
TANBA Sumio Hirosaki University, Faculty of Science and Technology, Associate Professor, 理工学部, 助教授 (70179927)
|
|---|
| Project Period (FY) |
2004 – 2005
|
| Keywords | wind erosion / satellite image / geometric correction / illumination correction / land cover clasification / spatial segmentation / generalized linear model / environmental factor |
|---|
| Research Abstract |
In this research, wind erosion areas in Kitakami highland of Iwate Prefecture were identified by using satellite images, and its relationship with environmental factors was analyzed. In addition, the effect of counter measures for wind erosion prevention were cholonologically analysed.
For quantitative analysis, the satellite images were carefully processed with geometric correction and illumination correction. As the optimization method for geometric correction using simulated relief image cannot be applied to wide area, we proposed a modified algorithm, which model the shift of scene center position as a linear regression function. By using the new algorithm, the whole Kitakami highland was covered by the satellite images seamlessly. For illumination correction, we applied the modified cosine method and obtained the images with less shading relief. We did not apply any special atomospheric correction because of high surface altitude of the study region.
The corrected satellite images w
… More
ere used for land cover classification : first the images were spatial segmented by using pyramid linking method, then the segments were clustered using K-Mean method, finally they were classified into five categories : (1)eroded soil, (2)bair soil, (3)grass, (4)forest *****ages were used for analyzing its relationship with following environmental factors : (a)surface elevation, (b)slope angle, (c)aspect angle, (d)openness to the west, (e)forest existence on the west, (d)solar radiation in spring. Each factor was analysied for its effect on the erosion rate, and all factors were incorporated into the generalized linear model. The analysis revealed that the forest existence was the most important factor, which prevent the west wind in winter.
There was some difficulties to identify the erosion area. For example, acute slope of more than 35 degree and the rocked summit of Mt.Hayatine looked like erosion area from the satellite image. These areas were carefully eliminated from the erosion category. Utilization of satellite images, however, was proved to be very effective to identify the erosion area over a wide region, and we found some erosion area which had not been identified so far. Less
|
Research Products
(12 results)
