|
遥感学报 2000
Analysis of Sensibility on Split-window Algorithms for Retrieving Land-surface Temperature
|
Abstract:
This paper presents two models for describing error of transition curve features based on error of points on the curve. The error nature of such a curve depends on variance covariance matrices of endpoints. The first error model is based on root mean square error in normal direction of the transition curve. The second model is based on error in the maximum direction of the curve. Theoretically, the error model of maximum direction is more rigorous for describing error of transition curve lines. The error model of normal direction is a simplified model of the previous one and is widely used in describing error of straight lines.