The Rectification of High Resolution Remote Sensing Satellite Imagery
4.2 Case 2
Case 2 refers to the image rectification with accurate control in various pattern of distribution. Control pattern were classified into 2 types.
The first type is for flat terrain control in which 16 planimetric control points were selected. Among them 9 points served as the control and the rest of 7 points served as the check points.
The second test refers to ortho-rectification. The correspondent 1/1,000 digital topographic map was served as the DTM. Figure 3 is the origin scene of Yuan-shan. Figure 4 illustrates a portion of the rectified image in which the digital topographic map is on top of it to provide a visual impression of the processing quality.
5 Discussion and Conclusions
All the results in this study are evaluated by the Root Mean Square Error (RMSE) value. Table 1 shows the control pattern and its error vectors using planimetric control. Table 2 shows the RMSE of the ortho-rectification process.
The RMSE of the control points are all less than 1 pixel. The RMSE of the check points are also under 1 pixel in the flat area. While in hilly terrain, the RMSE increase drastically. These results also coincide with the theory that hilly terrain tend to decrease accuracy of height measurement, which shows at point number 15 with tremendous amount of error.
It is clear that after geometric rectification using accurate controls, the IKONOS image can be utilized in large scale mapping and planning. The image in this paper is adopted freely from the Internet and with little knowledge about the satellite attitude during exposure. It is strongly recommended that with the help of orbital parameters, one can expect a reliable and accurate image available for detailed earth observation. The short revision period of this imaging system also provides the possibility of rapid map updating which in general is a time consume and costly task.
Table 1: The control pattern and error vectors using planimetric control
Table 2: The RMSE of control points after ortho-rectification
| Pt #
|
RMSE
|
Pt #
|
RMSE
|
Pt #
|
RMSE
|
Pt #
|
RMSE
|
| 1
|
0.28
|
10
|
0.23
|
4
|
1.4
|
9
|
3.17
|
| 2
|
0.6
|
11
|
0.77
|
6
|
0.76
|
12
|
4.57
|
| 3
|
0.25
|
13
|
0.36
|
7
|
0.43
|
14
|
0.61
|
| 5
|
0.28
|
16
|
0.32
|
8
|
0.58
|
15
|
13.28
|
6 Acknowledgement
The images used in this paper are credited to SPACEIMAGING.COM. Thanks to Jen-yu Han, I-hwei Du and Na-wen Tsai for their help in conducting the experiences in this article.
7 References
- Mayr, 1988, ISPRS
Congress preceeding, Commission IV, Kyoto,
Japan, pp.430-439.
- Lillesand and Kiefer,
2000, Remote Sensing and Image Interpretation,
pp.309-438.
- Toutin and Cheng, 2000,
Densifications of IKONOS, Earth Observation
Magazine, V9, No 7, pp.13-21
- Visser, 1980,
Orthophoto Production and Application, The ITC
Journal, 1980-4, pp.638-659.
- Wiesel, 1985, Digital Image Processing for Orthophoto Generation, Photogrammetria, 40,pp.69-76.
