Shotaro ABE
Development of snow coverage area distinction methodology from the terrestrial images considering lighting conditions on snow surface
Atsushi RIKIMARU, Kazuyoshi TAKAHASHI
The research on the snowfall that uses the remote sensing technology has been executed by the satellite observation and the ground observation image. For the ground observation, it becomes impossible for the image that there is a change in the lighting condition to distinguish the snow coverage area accurately. Therefore, this thesis purposed to examine the snow coverage area distinction methodology for considering the lighting condition on snow surface, and to attempt the improvement of the snow coverage area extraction accuracy.
It proposed limited part distinction methodology as a distinction methodology for improving the extraction accuracy of the shade snow coverage area in this thesis. This limited part distinction methodology distinguishes the snow coverage area by using the dynamic threshold that changes according to the lighting condition, and it comes to be able to extract the snow coverage area in the shade part.
To verify the utility of the limited part distinction methodology, the comparative verification with the snow coverage area distinction methodology in the past was executed. As for the limited part distinction methodology, the extraction accuracy of the shade snow coverage area improved by about 30%, and it was confirmed to be able to extract the snow coverage area even in the image with a different lighting condition by steady accuracy of 80% or more by the verification.
From the above-mentioned content, as for the limited part distinction methodology proposes with this thesis, can be said that the snow coverage area extraction accuracy improved more than the distinction methodology in the past.
Please use "previous botton" to return the previous page