Hình ảnh viễn thám ngày càng Được sử dụng cho bộ sưu tập các thông tin không gian. Một phát triển rộng rãi trong các hệ thống cảm biến đã xảy ra TRONG Thập kỷ trước, Kết quả là Cải thiện độ phân giải không gian, thời gian, và quang phổ. Các bộ sưu tập các dữ liệu bằng cách từ xa Nói chung cảm biến hiệu quả hơn IS và bằng quan sát trực tiếp rẻ hơn và đo lường trên mặt đất, mặc dù vẫn có chất lượng khác nhau. Thu thập dữ liệu bằng cảm biến May Be bị ảnh hưởng bởi. | 14 Use of the Bradley-Terry Model to Assess Uncertainty ỉn an Error Matrix from a Hierarchical Segmentation of an ASTER Image Alfred Stein Gerrit Gort and Arko Lucieer CONTENTS Concepts and The The BT Formulating and Testing a Case Study. .330 The Error Implementation in The BT The BT Model for Standardized Data . 336 Discussion . 337 Conclusions . 339 References . 339 Introduction Remotely sensed images are increasingly being used for collection of spatial information. A wide development in sensor systems has occurred during the last decades resulting in improved spatial temporal and spectral resolution. The collection of data by remote sensing is generally more efficient and cheaper than by direct observation and measurement on the ground although still of a varying quality. Data collected by sensors may be affected by atmospheric factors between sensors and the values reflected on the earth s surface local impurities on the earth s surface technical deficiencies of sensors and other factors. In addition only the reflection of the sensor s signal or of the sunlight on the earth s surface is being measured and no direct measurements are made. Consequently the quality of maps produced by remote sensing needs to be assessed 1 . Another major issue of interest concerns the ontological interpretation of remote-sensing imagery. The truth as it supposedly exists at the earth surface is due to what one wishes to see. This can be formalized in terms of formal ontologies although these may be subject to changes in time as well and also a subjective interpretation is often being given. 325 2008 by Taylor Francis Group LLC 326 Image Processing for Remote Sensing Image classification and segmentation are important concepts in this respect. A segmentation of an image yields segments of more or less .
