C Conferentia Proceedings
NSC2016-0057 Time Series Analysis

Using two-dimensional continuous wavelet transform to detect differences among primary forest, water bodies, clouds and cloud shadows on remote sensing images of an Amazon rain forest region

Cledenilson Mendonça de Souza1; Margarete Oliveira Domingues2; Marcos Adami3; Leonardo Deane Sá3

1 Universidade Federal do Amazonas (UFAM/ICSEZ); 2 INPE; 3 Centro Regional da Amazônia-Instituto Nacional de Pesquisas Espaciais

doi:10.20906/CPS/NSC2016-0057

Resumo

One issue in practical uses of remote sensing im- agery is to map clouds and cloud shadows. Often the clouds can be confused with snow, sand or straw cover. Moreover, the shadows of the clouds can be confused with water bod- ies. In this way, we are using Two-Dimensional Continu- ous Wavelet Transform (CWT2D) in order to identify differ- ences between primary forest, water bodies (lakes and rivers) from clouds their shadows. The site is located at the Uatumã Reserve (USDR) in Central Amazonia. We use CWT2D to detect per scale differences among patterns of the surface texture. The results show the existence of statistically sig- nificant differences among the distinct classes of areas men- tioned above

Palavras-chave: Time Series Analysis; infrared images; cloud and cloud shadow and water body detection; wavelet transform

Como citar

Cledenilson Mendonça de Souza; Margarete Oliveira Domingues; Marcos Adami; Leonardo Deane Sá. “Using two-dimensional continuous wavelet transform to detect differences among primary forest, water bodies, clouds and cloud shadows on remote sensing images of an Amazon rain forest region”. 6th International Conference on Nonlinear Science and Complexity. NSC2016. 2016. DOI: 10.20906/CPS/NSC2016-0057