航空辐射测量与机器学习算法揭示土壤质地
摘要：意大利Mezzano低地是一个面积为189平方公里的农业平原，我们通过一次专门的航空伽马射线光谱测量调查，研究了放射元素与土壤质地之间的关系。我们利用K和Th的丰度，采用多种方法来获取粘土和砂的含量。采用线性（简单和多元）和非线性（具有深度神经网络的机器学习算法）预测模型，使用1:50,000比例尺的土壤质地地图进行训练和测试。这些方法的比较表明，非线性模型在土壤质地分数的预测中引入了显著改进。预测的粘土和砂含量地图与区域土壤地图进行了比较。尽管宏观结构均存在，但航空伽马射线数据可以揭示出更精细的特征。高粘土含量的地图区域与Mezzano低地古代的Etruscan和Roman时期的古河道相符，这一点可以通过历史地图的水文设置和研究区域的地貌特征来确认。
作者：Andrea Maino (1 and 2), Matteo Alberi (1 and 2), Emiliano Anceschi (3), Enrico Chiarelli (1 and 2), Luca Cicala (4), Tommaso Colonna (5), Mario De Cesare (4 and 6 and 7), Enrico Guastaldi (5), Nicola Lopane (1 and 5), Fabio Mantovani (1 and 2), Maurizio Marcialis (8), Nicola Martini (9), Michele Montuschi (1 and 2), Silvia Piccioli (9), Kassandra Giulia Cristina Raptis (1 and 2), Antonio Russo (3), Filippo Semenza (1 and 2), Virginia Strati (1 and 2) ((1) Department of Physics and Earth Sciences, University of Ferrara, Ferrara, Italy, (2) INFN Ferrara Section, Ferrara, Italy, (3) Gruppo Filippetti Sede Falconara Marittima, Falconara Marittima, Ancona, Italy, (4) CIRA, Italian Aerospace Research Centre, Capua, Caserta, Italy, (5) GeoExplorer Impresa Sociale S.r.l., Arezzo, Italy, (6) Department of Mathematics and Physics, University of Campania Luigi Vanvitelli, Caserta, Italy, (7) INFN Napoli Section, Complesso Universitario di Monte Sant'Angelo, Napoli, Italy, (8) Comacchio, Ferrara, Italy, (9) Le Due Valli S.r.l., Ostellato, Ferrara, Italy)
论文ID：2208.06625
分类：Geophysics
分类简称：physics.geo-ph
提交时间：2022-08-16