The significance of spatial resolution: Identifying forest cover from satellite data

Abstract

The objective of this study was to determine if decreasing pixel size increases overall accuracy with which forest species can be separated using satellite data. Classification accuracy achieved with Landsat Thematic Mapper (TM) channels 2,3, and 4 were compared with result from SPOT-XS. Reference data prepared from enlarged prints of aerial photographs and field checks included 19 cover types, nine of which were coniferous. Results with SPOT-XS alone yielded an overall accuracy of 70.1 percent, and with Landsat TM 57.1 percent at species level. Overall accuracy increased by 3 percent when either SPOT-XS or Landsat TM data were integrated with SPOT-PAN data.