Abstract
It is well known that change detection and categorization are of limited accuracy due to various man-machine factors. However, it is a complicated matter to estimate accuracy levels in land cover change information derived from remote sensing data due to spatio-temporal interdependence (and inhomogeneity) in error occurrences, non-trivial extension of accuracy measures from single-date classifications to change detection, the often inadequacy of ground data, and other limitations. This paper seeks to provide a synthesis of some of the potentially useful methods for estimating accuracy in land cover change information. Geostatistical approaches will be valuable tools, though with computational and modeling costs, for providing accuracy-related information as they can incorporate spatio-temporal interdependence in estimating pixel-specific probability of correct classification.
| Original language | English |
|---|---|
| State | Published - 8 Jul 2014 |
| Event | 11th International Symposium on Spatial Accuracy Assessment in Natural Resources and Environmental Sciences, Accuracy 2014 - Duration: 8 Jul 2014 → 11 Jul 2014 |
Conference
| Conference | 11th International Symposium on Spatial Accuracy Assessment in Natural Resources and Environmental Sciences, Accuracy 2014 |
|---|---|
| Period | 8/07/14 → 11/07/14 |
Indexed by
- EI
Fingerprint
Dive into the research topics of 'Estimation of probabilistic accuracy measures in remotely sensed land cover change information'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver