Home Articles Multi-Sensor Image Fusion for Urban Expandsion of Chiang Mai City

Multi-Sensor Image Fusion for Urban Expandsion of Chiang Mai City

Mr.Ekkarat Pricharchon, Ms. Supapis Polngam
Geo-Informatics and Space Technology Development Agency (Public Organization)
196 Phahonyothin Road,Chatuchak, Bangkok 10900 Thailand
Email: [email protected]

1. Introduction
An operational earth observation, optical sensors have proven an efficient tool for many applications. Imagery acquired by them offer various wavelengths and reveal specific information about ground features. Applications of satellite remote sensing have been accomplished in several fields. The Geo-Informatics and Space Technology Development Agency use satellite data for the project. Multi-sensory data provides complementary information about scene. In consequence, image fusion improves the interpretability of the images, recently, higher quality satellite images are available than ever.

2. Study area


Fig. 1. Chiang Mai area in northern of Thailand.
2. Data Used

2.1 Satellite Data
SPOT-1 Panchromatic 1988
LANDSAT-5/TM 1988
LANDSAT-7/ETM 2002
IRS-1D Panchromatic 2002
Table 1. Landsat-5/ 7 image
Band Spectral range(mm) Resolution(m) 3 0.63-0.69 30 4 0.75-0.90 30 5 1.55-1.75 30


Fig. 2. Landsat image color composite 453:RGB, 30m resolution.
Table 2. SPOT-1 image
Band Spectral range(mm) Resolution(m) Pan 0.51-0.73 10


Fig. 3. SPOT-1 image panchromatic, 10m resolution.
Table 3. IRS-1D image
Band Spectral range(mm) Resolution(m) Pan 0.53-0.75 5.8

Fig. 4. IRS-1D image panchromatic, 5.8m resolution.
2.2 Other related Information
Topographic maps which were produced by the Royal Thai Survey Department and an existing land use map from Land Development Department at a scale 1:50,000 were used for supporting image interpretation.

3.Methedology
The simple image pre-processing was carried out including image enhancement, the higher resolution image is used as the reference to which the lower resolution image is geometrically registered and Pan-sharpening (the combination of SPOT panchromatic image data, having a spatial resolution of 10m, with Landsat-5 Thematic Mapper images, having a spatial resolution of 30m in year of 1988 and the combination of IRS-1D panchromatic image data, having a spatial resolution of 5.8m, with Landsat-7 Enhance Thematic Mapper images, having a spatial resolution of 30m in year of 2002.). At this stage, Pan-sharpening images can provide a hybrid image having a good spatial detail for the visual interpretation of 1: 50,000 scale, and also ground truth data which were collected by using GPS tracking in 2002 can be used for this step. Urban land use map of 1988 and 2002 were obtained by using GIS to detect urban land use change.


Fig. 5. Pan-sharpening image from Landsat-7 with IRS-1D
4. Results
The pan-sharpened image as shown in Figure 5 provided a better interpretation capability for urban land use mapping. Land Use categories were derived from visual interpretation. The image interpretation keys used to identify includes tone, texture, pattern, shape and association. The land use types were divided as 27 class, after that group for reduce diversity of class into 5 classes in Table 4.The land use of 1988 and 2002 as will as the land use detection are shown in Figures 6-8.

Table 4 Land use and land use change of Chiang Mai during1988 – 2002
LAND USE Area (sq.km.) Land use change 1988 2002 Sq.km. % Agriculture 298.3 110.79 -187.51 -62.86 Forest land 13.35 3.11 -10.24 -76.7 Water body 2.49 2.89 0.4 +16.06 Urban& built up land 94.4 162.73 68.33 +72.38 Miscellaneous land 1.52 130.5 128.98 +8485.5


Fig. 6-8. Land use 1988, 2002 and change detection
5. Conclusion
The best identification of urban land use type with pan-sharpened image of IRS Panchromatic and LANDSAT-7+ETM is achieved by visual interpretation using reflectance in term of tone and color, texture, pattern, shape and association. In addition, understandings of local cultural practices, knowledge of topography and ground information are essential to make an accurate urban land use map.