作者: A. M. Genú , J. A. M. Demattê
DOI: 10.1080/01431160600568926
关键词: Soil water 、 Radiometry 、 Soil classification 、 Soil test 、 Spectral data 、 Environmental science 、 Reflectivity 、 Regression analysis 、 Linear regression 、 Mineralogy
摘要: Remote sensing involves huge data bases of spectral reflectance in developing a new soil analysis method. The objective this work was to use determine the contents physical, chemical, and mineralogical attributes 9 classes with sensor laboratory. To that end, database organized containing information on 3,300 samples their respective analyses radiometry, from Brazilian states Sao Paulo Parana. With data, multiple regression equations were obtained for 20 attributes. physical presented had, most part, coefficients determination higher than 0.70, while chemical value was, general, smaller 0.50. Attributes whose models R2 > 0.50 tested unknown did not take part generation model. model‐estimated content values those then compared determined labora...