Reserch The System Environmental Change Extraction
using Satellite Data
2.3 Make the manual of interpretation for microwave imagery
-
Manual of interpretation for microwave imagery
In order fill up manual of interpretation for SAR (Synthetic Aperture Radar) imagery in phase I (Hiromichi, 1995), we obtained additional ground truth with filed reconnaissance survey using RADARSAT imagery of high resolution. And we argued with staff of DLD about result of field reconnaissance survey and kmicrowave imagery.
- Specification of manual
The manul of interpretation for microwave imagery has LANDSAT-TM imagery, RADARSAT imagery and composite imagery of both satellite for each category. Still more, it put on field photograph and each category. Each classification category are 9 items same as(1).
3) Result of Research
3.1 Develop the method of environmental change extraction using satellite data
The result of land cover classification on two different times is shown in figure 3. The X axis is category. The Y axis is land cover proportion on two different times (1987,1995). The large attention points were decrease of field, increase of pineapple and urban as a large change. We could guess that turning into housing and spreading agricultural land had changed from the area of field and waterland. The influence of the land improvement around the large reserved water on north-west area could be guessed from the result of classification too. In the future, the north-west area will change because the purpose of the large reserved water was industrial water.
Figure 3. Land cover of each items
The change process flow chart of land cover is shown in figure 4. The selected items of change are large change proportion so that it represents the improvement of land and urbanization, were referred to opinion staff of DLD. The figure above arrow was shown the area (km x km) of change. Even if only two different times data were used to classify and
extract the change, we could guess rough process of land cover in this area. For example change process was guessed as follows.
Figure 4. Change flow of land cover
Field ® pineapple ® mix-orchard
® urban. Or field ® urban.
Pineapple and mix-orchard, rubber and mix-orchard were used land by reciprocal relation. And we could guessed that increase proportion of field
® pineapple, decrease proportion of mix-orchard. The change of land cover has close relation with soil, which was difficult to change form of land use. In this area, it has characteristic that land use depend on soil of geographical. The plantation was in order of gentle slope, which was rice field
® mix-orchard
® pineapple, rubber. So we could understand that the area of gentle slope spread due to improve land.
3.2 Develop the method of environmental change analysis using GIS
The land cover map that was overlaid with geographical data which were road, river and controur line, executed geometric correction is shown in figure 5. We could visually understand land cover due to be overlaid with geographical data. The land cover was divided from road and river. Particularly, we could understand that a lot of urban area existed around the road and river, pineapple existed more elevated than mix-orchard rubber was located along divided the river.
Figure 5. 95 Land cover map
The overlaid change extraction with geographical data was shown in figure 6. We could visually understood that the parts of change extraction had close relation with geographical data. And it was clearly understood that a lot of change of urban was around the road, changed a lot of pineapple in north-west area and did not change in east area.
Figure 6. Composite map of land cover change
The result of spatial analysis from road and river using buffering method was shown in table 1.
Table 1. The proportion of change items from road and river using buffering
| Items of change |
Distance from road |
Total |
| 100m |
300m |
600m |
1km |
Over1km |
| Pineapple |
9.4 |
20.7 |
22.3 |
20.0 |
27.5 |
100 |
| Reserved water |
8.9 |
20.2 |
26.2 |
23.6 |
21.1 |
100 |
| Rubber |
4.9 |
9.9 |
14.7 |
17.6 |
52.9 |
100 |
| Wateland |
1.7 |
4.9 |
7.1 |
11.5 |
74.7 |
100 |
| mix-orchard |
6.1 |
13.5 |
16.5 |
19.6 |
44.3 |
100 |
| Urban |
22.4 |
24.1 |
22.2 |
15.1 |
16.2 |
100 |
| Items of change |
Distance from road |
Total |
| 100m |
300m |
600m |
1km |
Over1km |
| Pineapple |
9.2 |
32.2 |
36.4 |
17.2 |
5.4 |
100 |
| Reserved water |
29.6 |
34.8 |
21.7 |
9.5 |
4.4 |
100 |
| Rubber |
13.7 |
22.7 |
27.4 |
19.2 |
17.0 |
100 |
| Wateland |
9.0 |
21.7 |
28.7 |
23.5 |
17.1 |
100 |
| mix-orchard |
10.5 |
22.4 |
26.7 |
22.3 |
18.1 |
100 |
| Urban |
17.8 |
28.8 |
25.1 |
15.2 |
13.0 |
100 |
We could executed quantitative analysis using buffering method. For example, change of urban was about 50% within 300m within 300m form. Change of mix-orchard, rubber and wasteland were not about 50% within 1km from road. Change of all items were about 60% into 600m from river. We could understood that a lot of change of pineapple is elevation from 50m to 70m, a lot of change of mix-orchard was elevation from 20m to 50m and a lot of change of urban was elevation from 0m to 20m.
