| Abstract | The effective integration of remote sensing data and geographic information is essential if one is to able to update a Geographic Information System (GIS) in a timely fashion. For an inventory for
forest management, estimates of species, stand density, crown enclosure, age and site quality are needed. The Canada Centre for Remote Sensing (CCRS) and the British Columbia Ministry of Forests (BCMOF) are conducting experiments integrating remote
sensing and geographic information on remote sensing data from LANDSAT Thematic Mapper (TM) and the airborne imager, MEIS. The acquired data are geometrically corrected, including corrections for topographic relief. GIS forest cover files, describing
historical conditions over the area, are analyzed in order to develop likely interpretations for objects detected in the imagery. Initial work has established that clustering TM and MEIS data within polygons with the same GIS class label can lead
to the identification of crown closure and stand density differences within GIS polygons. Thus, in this case, remote sensing was able to provide information at finer detail than that currently included in the inventory derived from interpretation of
aerial photography. Segmentation of the GIS files and the image data permit us to create objects made up of multiple segments. The historical data support the interpretation of the image segments and enable one to detect any inconsistencies in the
polygon labels or boundaries. Changes in forest cover as a result of logging operations are also identified. This presentation will focus on the use of remotely-sensed data for forest inventory, at a range of spatial and spectral resolutions, and
ground information to understand the underlying physical causes of the different cluster found within GIS polygons from the image data. |