@article{oai:repository.naro.go.jp:00008806,
 author = {渡辺, 利通 and WATANABE, Toshimichi},
 journal = {農業環境技術研究所資料, Miscellaneous Publication of the National Institute of Agro-Environmental Science},
 month = {Mar},
 note = {水稲の収量推定におけるリモートセンシング技術の利用の可能性について検討することを目的として，ランドサットMSSデータを用いた福岡県，佐賀県の市町村別の水稲の10アール当たり収量の推定を試みた。その結果，MSSのバンドデータおよびそれらを組み合わせた合成変数を用いた重回帰分析により，重相関係数が0.6499の重回帰式がえられた。この推定式によりこれらの対象地域における10アール当たり収量の概略的な分布を推定することができた。しかし収量推定におけるリモートセンシングの実用的な利用についてはさらに研究の積み重ねが必要である。, Author examined to estimate the rice yield per unit area and to classify paddy field from LANDSAT MSS data using ARSAS. Tested area was Fukuoka and Saga prefectures, and data used were LANDSAT MSS data (1979.10.9, path 121-row37,"Kitakyusyu" scene),1/50000 map and Crop Statistics of 1979. (1) 141 fields cultivating rice crop were extracted from 1/50000 maps. Longitude and latitude of these points was translated to line and pixel number on LANDSAT MSS image data through the UTM coordinate system.Then MSS band data of any points were read and analyzed by multiple regression analysis using yield per 10 ares as the dependent variable and MSS band data and 106 variables composed of some band data as independent variables. From multiple regression analysis, it was shown that yield per 10 ares could be roughly estimated from some composite variables with band data. At this case multiple correlation coefficient (R) was O.6499. But it was suggested that actual application of remote sensing to yield estimation on other years or regions was difncult, because this equation had low contributing rate to explain the variation of yield per 10 ares and biological or agricultural sense of this equation were not definite. Therefore more studies over years, regions, cultivars or seasons are necessary for more accurate estimation of yield of rice crop. (2) Paddy field area was distinguished by supervised classincation method using maximum likelihood criterion from same data. It was shown that paddy field was distinguished from other categories as mountains, water or urban by setting suitable training areas. Using same data, unsupervised classincation method got the similar categorical distribution.},
 pages = {21--35},
 title = {ランドサットMSSデータを用いた九州地方における水稲収量推定の試み},
 volume = {4},
 year = {1988},
 yomi = {ワタナベ, トシミチ}
}