CLOUD EXTRACTION FROM POLAR SATELLITE DATA USING MODIFIED MAHALANOBIS CLASSIFIER
In the polar region, it is difficult to discriminate between clouds and ground surface from satellite visible or infrared data, because of the high albedo and low surface temperature of snow and ice cover. In addition, since the latitude is high, visible channels cannot be used during winter. In thi...
Main Authors: | , , , , , , , |
---|---|
Format: | Report |
Language: | English |
Published: |
Faculty of Engineering, Kanazawa University
1998
|
Subjects: | |
Online Access: | https://nipr.repo.nii.ac.jp/?action=repository_uri&item_id=2871 http://id.nii.ac.jp/1291/00002871/ https://nipr.repo.nii.ac.jp/?action=repository_action_common_download&item_id=2871&item_no=1&attribute_id=18&file_no=1 |
Summary: | In the polar region, it is difficult to discriminate between clouds and ground surface from satellite visible or infrared data, because of the high albedo and low surface temperature of snow and ice cover. In addition, since the latitude is high, visible channels cannot be used during winter. In this paper, a method to extract clouds using only NOAA/AVHRR channel 4 is proposed. Using geographical information, the AVHRR image was first segmented into two regions : sea and land. Afterward, cloud extraction was performed for each region separately by minimum distance classifier using image features. To improve the error rate of the classification, we apply thresholds to the discriminant function used by the Mahalanobis classifier. |
---|