Loading…
NECEC 2017 has ended
View analytic
Wednesday, November 15 • 11:20 - 11:40
A Novel Method for Classification of Complicated Land Covers Using Remote Sensing Techniques

Sign up or log in to save this to your schedule and see who's attending!

Land cover classification is a common application of Remote Sensing, which often yields desirable results. However, this task becomes quite challenging when the target classes are spectrally similar. In this paper, a method is proposed which attempts to alleviate this issue. In the proposed method, not all the objects are given a label at the same time, but the objects belonging to each class are delineated in different steps and each step is associated with a separate feature selection and classification. The suggested approach was applied to the Avalon peninsula, located in Newfoundland and Labrador, Canada. The obtained average producer and user accuracies of wetlands was 79% and 77% , respectively, which was 17% and--% percent superior to those of classification using the typical method. Therefore, the proposed approach can be potentially applied to distinguish between the classes of any complicated land cover.


Wednesday November 15, 2017 11:20 - 11:40
Conception Bay South 180 Portugal Cove Road, St. John's, NL, Canada