| Title | A comparison of different remotely sensed data for classifying bedrock types in Canada's Arctic: Application of the Robust Classification Method and Random Forests |
| |
| Author | Harris, J R; He, J X; Rainbird, R ; Behnia, P |
| Source | Geological Association of Canada-Mineralogical Association of Canada, Joint Annual Meeting, Abstracts Volume vol. 37, 2014 p. 115  Open Access |
| Links | Online - En ligne
|
| Image |  |
| Year | 2014 |
| Alt Series | Earth Sciences Sector, Contribution Series 20130439 |
| Publisher | Geological Association of Canada |
| Meeting | Geological Association of Canada - Mineralogical Association of Canada Joint Annual Meeting; Fredericton; CA; May 21-23, 2014 |
| Document | serial |
| Lang. | English |
| Media | paper; on-line; digital |
| File format | pdf |
| Subjects | geophysics; arctic geology; climate, arctic; remote sensing; vegetation |
| Program | GEM: Geo-mapping for Energy and Minerals GEM Tri-Territorial Information management & Databases (Remote Predictive Mapping / Mineral Resource
Assessment) |
| Released | 2014 01 01 |
| Abstract | Over the past three decades, the increasing availability of space - borne sensors imaging the Earth's surface using increasingly higher spatial and spectral resolutions has evolved geologic remote
sensing from being primarily a qualitative discipline to a quantitative discipline based on the computer analysis of digital images. The Geological Survey of Canada under the Remote Predictive Mapping (RPM) project part of the Geo - mapping for
Energy and Minerals (GEM) program, Natural Resources Canada, has the mandate to produce up - to - date geoscience maps of Canada's territory north of 600. Classification of remotely sensed data is a well - known and common image processing
application that has been used since the early 1970's concomitant with the launch of the first LANDSAT (ERTS) earth observational satellite. In this study we apply supervised classification using a new algorithm known as the Robust Classification
Method (RCM) and a Random Forest (RF) classifier to a variety of remotely sensed data including LAND SAT - 7, LANDSAT - 8, SPOT - , ASTER and airborne magnetic imagery producing predictions (classifications) of primarily bedrock lithology and
quaternary cover in central Victoria Island, Northwest Territories. We compare and contrast these different data types and evaluate how well they classify various lithologies and surficial materials using confusion analysis (confusion matrices) as
well as comparing the generalized classifications with the newly produced geology map of the study area. In addition we propose some new ensemble classification methods that leverage the best characteristics of all remotely sensed data used for
classification. Both RCM and RF provide good classification results. However, RF provides the highest classification accuracy because it used all 43 of the raw and derived bands from all the remotely sensed data. The ensemble classifications based on
the generalized training dataset showed the best agreement with the new geology map for the study area. |
| Summary | (Plain Language Summary, not published)
This abstract presents results dealing with the classification of various remotely sensed data for lithological mapping in Canada's Arctic (Victoria
Island). The results indicate that in a geological environment where rocks are well exposed, classification can contribute signicantly to the geologic mapping process. |
| GEOSCAN ID | 293581 |
|
|