@techreport{oai:jaxa.repo.nii.ac.jp:00045112,
 author = {松本, 甲太郎 and 中, 正夫 and 山本, 浩通 and MATSUMOTO, Kotaro and NAKA, Masao and YAMAMOTO, Hiromichi},
 month = {Mar},
 note = {The clustering technique is important as the unsupervised classifier in the realization of an automated satellite image analysis system. But current cluster analysis methods are not suitable to deal with huge, multi-dimensional satellite images because of the dependency on initial partitions and the tremendous computational time required. In this paper, to analize a huge satellite image, we will describe two new algorithms, i.e., Multi-Layered Hashing Scheme and Histogram Mode Method. The first one is for the construction of a multi-dimensional histogram from the huge satellite image data. And the second one is for deriving natural clusters from the histogram using the basic idea above. The algorithms were implemented on the large scale pipeline computer (FACOM 230/75 APU), and have been successfully used. Some results are shown in the paper., 資料番号: NALTR0854000, レポート番号: NAL TR-854},
 title = {大規模衛星画像クラスタ解析の研究-ヒストグラム・モード法-},
 year = {1985}
}