DEVELOPMENT OF THE NEW ALGORITHM FOR MANGROVE CLASSIFICATION

I WAYAN NUARSA (Udayana, Bali)
I WAYAN SANDI ADNYANA (Udayana, Bali)
YASUHIRO SUGIMORI (CeRES, Chiba University, Japan)
SUSUMU KANNO (CeRES, Chiba University, Japan)
FUMIHIKO NISHIO (CeRES, Chiba University, Japan)

International Journal of Remote Sensing and Earth Sciences (IJReSES)
Vol 2(2005)

The objective of the study is to develop the algorithm for mangrove classification and density. Regression and correlation analysis was used to perform the equation. CE1 = (0.663*Band 3) + (0.l55 *Band 4) - (l.4*Band 5) + 0.995 And CE2 = 36 * Band 4 + 6*Band 5 + Band 3 were two formula that have been used to classify the mangrove. The object will be classified as mangrove when the value of CE1 is between -31.439 and 0.888, and value of CE2 is greater than or equal to 2000. On the other hand, density of the mangrove was expressed as DE = (2 * Band 4)/(Band 1+Band 3). Mangrove classification result in this study was similar to those of the existing methods. Statistical approach in this study generally gives the higher result tendency than other methods. Key words: Mangrove, Landsat ETM+, Empirical Model, Image Classification