| タイトル | Classification results using spacially correlated Landsat data |
| 著者（英） | Tubbs, J. D. |
| 著者所属（英） | Arkansas Univ. |
| 発行日 | 1979-01-01 |
| 言語 | eng |
| 内容記述 | Tubbs and Coberly (1978) demonstrated that Landsat multispectral scanner data are not independent random observations, but, are in fact highly correlated. They also demonstrated that the correlation structure for the data is similar to that of a stationary autoregressive process of order one. This paper investigates the effect that serially correlated training data have upon both the estimation of parameters and the classification problem. Results are included for both the Bayesian and maximum likelihood classification procedures. |
| NASA分類 | EARTH RESOURCES AND REMOTE SENSING |
| レポートNO | 80A22481 |
| 権利 | Copyright |
| URI | https://repository.exst.jaxa.jp/dspace/handle/a-is/423086 |
|
