Utilizacao de dados do SAR JERS-1 em modelos preditivos de umidade do solo e coeficiente de retroespalhamento / Utilization of JERS-1 SAR data for soil moisture and backscatter coefficiente retrieval models
AUTOR(ES)
Orlando Alves Maximo
DATA DE PUBLICAÇÃO
1997
RESUMO
The objective of this study is to evaluate the feasibility on the use of JERS-1 SAR data for soil moisture retrieval models based on radar images, and to compare the results with field data. Two models that relate the backscatter coefficient with soil parameters such as soil moisture and roughness were used: Oh et al. (1992, 1994) and Dubois et al. (1995). A 16 bits amplitude image was calibrated to obtain the backscatter coefficient. In order to minimize the influence of speckle, the calibration was taken using the average of 16 pixels. The model from Dubois et al. reaches a higher correlation with calibrated JERS-1 SAR data than Oh s model. The latter could not be inverted because it includes a non-invertibte function, namely the modulus of a complex number. The Dubois model was inverted, but required the insertion of estimated values of roughness to be used with monochannel SAR data, such as JERS-1. The results were very satisfactory. The inversion model is quite sensitive to the estimation of roughness values. For bare soils, with an error of 5 mm of height (RMS), the correlation between computed and measured data became deteriorated..
ASSUNTO(S)
umidade do solo jers-1 modelos preditivos sar coeficiente de retroespalhamento radar de abertura sintética
ACESSO AO ARTIGO
http://urlib.net/sid.inpe.br/iris@1912/2005/07.20.11.05.43Documentos Relacionados
- Utilizacao de dados do SAR JERS-1 em modelos preditivos de umidade do solo e coeficiente de retroespalhamento
- Qualidade geometrica de imagens JERS-1 (SAR)e Radarsat (modos S2 e S3)em ambiente de floresta tropical (Rondonia)
- A temporal behaviour analysis of the land cover l band backscatter in the Tapajós national forest, state of Pará, by using jers-1 images
- Transition models for binary data
- Models for count data with applications