|
|
|
Regional/Global Environment
|
Land Cover Classification of Asia Region Using NOAAAVHRR 1-KM Dataset
6. Classification
In this study, we based on the phonological information from the ratio between monthly land surface temperature Ts and multi-temporal NDVI data to perform land cover classification. Ground truth data itself remains in the final classified result. The other part is Classified by decision tree method.
Since decision tree method is extracted from clustering result, cluster analysis is necessary to explain at first.
6.1 Unsupervised classification
In this study, the initial segmentation of ratio between Ts and NDVI composites into seasonal greenness classes is performed using minimum distance unsupervised clustering. For each variable, an interactive self-organizing clustering routing based on minimum spectral distance criteria (ISODATA) was run several times on selected PC's, with different combinations of ISODATA parameters. The final number of clusters in the result was 100. Since the number of months retained per year varied from six to ten, there are seven months from April 1992 applied for clustering.
6.2 Preliminary labeling
Based on the ground truth data, the purpose of preliminary labeling is to provide a general understanding of the characteristics of each cluster, and to determine which classes have two or more disparate land cover classes represented within their spatial distribution (Dr. Bradleg C. Reed, 1996). Preliminary labeling involved inspecting the spatial patterns and spectral or multi-temporal statistics of each class. Comprising each class to reference data and decisions concerning land cover types.
6.3 Postclassification stratification
Postclassification stratification is used to separate classes containing two or more disparate land cover types. In this study, about 60 percent of the clusters represent multiple land cover types. Most of these types of problems are the result of spectral similarities between evergreen and deciduous forest, and natural and agricultural grassland. These problems can be usually solved by developing criteria bases on the relationship between the confused seasonal greenness classes and selected ancillary datasets (Reed, 1996). In this study, the following ancillary datasets were analyzed:
-
Digital Elevation Model (DEM)
-
Maximum NDVI: the maximum monthly NDVI value in twelve months
-
Minimum NDVI: the minimum monthly NDVI value in twelve months.
There are two tasks involved in postclassification step. In order to separate the classes which have multiple land cover types, the first is to determine the ancillary variables and preliminary decision rules, and the second tasks is to implement and refine the decision rules (Brown, 1993). In this phase, the initial criteria are interactively trained, refined, and finally used to modify the original class.
Table 1. Land cover classification system(Tateishi. R & Wen. C.G. 1997)
7. Conclusion
A land cover map of the whole Asia was produced based on phonological information from ratio between land surface temperature Ts and multi-temporal NDVI using NOAAVHRR 1-km dataset.
Reference
-
BRADLEY, C. R., 1996. The U.S. Geological Survey's Global Land Cover Characterization Project. Photogrammetric Engineering and Remote-Sensing.
-
BROWN, J. F., LOVERLAND, T. R., MERCHANT, J. W., REED, B. C., 1993. Using Multisoruce Data In Global Land Cover Characterization. Photogrammetric Engineering and Remote Sensing, 59, 977-987.
-
HOLBEN, B. N., 1986. Characteristics of maximum-value composite image from temporal AVHRR data. International Journal of Remote Sensing. 7, 1417-1434.
-
KERBER, A. G, and SCHUTT, J. B., 1986. Utility of AVHRR channel 3 and 4 in land cover mapping. Photogrammetric Engineering and Remote Sensing. 52, 1877-1883.
-
LAMBIN, E. F., and EHRLICH, D., 1995. Combining vegetation indices and surface temperature for land cover mapping at broad spatial scales. International Journal of Remote Sensing. 16, 573-579.
-
LOVERLAND, D., T. R. MERCHANT, J. W., OHLEAN, D. O., and BROWN, J., 1991. Development of a land cover database for conterminous U. S. Photogrammetric Engineering and Remote Sensing, 57, 1453-1463.
-
RYUTARO, TATEISHI, and CHENG-GANG, WEN, 1997. Global Four Minute Land Cover Data Set. Journal of the Japan Society Photogrammetry and Remote Sensing, 36
-
PRICE, J. C., 1984. Land surface temperature measurements from the split-windows channels of the NOAA-7 Advanced Very High Resolution Radiometer. Journal of Geophysical Research. 89, 7231-7237.
|
|
|
|
|
|
|
