Abstract
Sound forest management requires accurate forest maps at an appropriate scale. Forest cover data developed at a national scale may be too coarse for forest management at a local level. We demonstrated a two-stage unsupervised classification, integrating Continuous Forest Inventory (CFI) data and Landsat imageries, to classify forest types for Indiana State Forests (ISF) and 8-km surrounding areas. In the first stage, an automatic unsupervised classification assisted by CFI data was applied in ISF. In the second stage, the resultant forest cover information from the first stage was used to expand the classification area into the 8-km surrounding areas. Splitting the classification procedure into two stages made it possible to expand the classification area beyond the coverage of the CFI data. This data-aided unsupervised classification approach increased the repeatability of forest mapping. The resultant map contains five forest types: conifer, conifer-hardwood, maple, mixed hardwood, and oak-hickory forests. The overall accuracy was 81.9%, and the total disagreement was 0.176. The accuracies of conifer, conifer-hardwood, maple, mixed hardwood, and oak-hickory forests were 81.6, 63.4, 75.0, 33.3, and 90%, respectively. This forest mapping technique is suitable for automated mapping of forest areas where extensive plot data are available.
Highlights
Remote sensing is an effective technology to extract and map spatially explicit information of land use and cover at different scales.[1,2] Forest cover maps are one of the products made using remotely sensed data and are essential for forest management, habitat monitoring, and biodiversity analysis.[3,4,5,6] General-purpose land cover maps developed at national and global scales usually describe forest information at a coarse level
The National Land Cover Database (NLCD)[7] for the United States contains three forest types: deciduous, evergreen, and mixed forests, a level-II classification defined by the United State Geological Survey (USGS).[8]
To separate oak-hickory forest from other forest types in stage 1, we started with 100 spectral clusters using unsupervised classification and increased the number of the clusters by a 50-cluster interval.[26]
Summary
Remote sensing is an effective technology to extract and map spatially explicit information of land use and cover at different scales.[1,2] Forest cover maps are one of the products made using remotely sensed data and are essential for forest management, habitat monitoring, and biodiversity analysis.[3,4,5,6] General-purpose land cover maps developed at national and global scales usually describe forest information at a coarse level. Fine-level forest cover maps with low classification accuracy are no better than coarse-level forest cover maps
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
