Python语言在物探数据清洗中的探索  

The application of Python in geophysical data cleaning

作　　者：谢明宏[1] 邱崇涛[1] 祁程[1] 张伟[1] 何昕欣 XIE Minghong;QIU Chongtao;QI Cheng;ZHANG Wei;HE Xinxin(Airborne Survey and Remote Sensing Center of Nuclear Industry, CNNC Key Laboratory for Geophysical Exploration Technology Center of Uranium Resource, Shijiazhuang 050002, China)

机构地区：[1]核工业航测遥感中心铀资源地球物理勘查技术中心重点实验室,石家庄050002

出　　处：《物探化探计算技术》2020年第6期814-822,共9页Computing Techniques For Geophysical and Geochemical Exploration

摘　　要：从物探数据采集到最终处理结果的可视化过程中,数据前期准备是极为重要的一环,需花费大量时间和精力。以洁净数据概念为基础,结合常见的物探测量数据特点,归纳总结了不洁净数据(集)常见类型,剖析了洁净数据与不洁净数据间内在关系,对物探数据(集)构建原则进行了初建。利用当今最为流行的科学计算语言Python,对常见数据清洗方法进行了列举。通过Python实例代码,展示了Pandas、Numpy、Spicy和Matplotlib等第三方库的强大功能,表明了规范的数据(集)结构配以Python语言可达到数据快速清洗的目的,从而大大节省数据预处理时间,为后续数据处理和资料解释提借鉴。From geophysical data acquisition to the final processing results and data visualization,data preparation is an extremely important part which takes a lot of time and effort.Based on the concept of tidy data by Hadely Wickham,combining with the characteristics of common geophysical data,1)the common types of messy data(dataset)are summarized;2)the inner relationship between tidy data and messy data are analyzed;and 3)the constructing principle of geophysical data(dataset)are built initially.Some data cleaning methods are exampled using Python,the current most popular scientific computing language.The Python codes demonstrate the power of third-party libraries such as Pandas,Numpy,Spicy,and Matplotlib,indicating that both standard data(dataset)structure and Python language can achieve rapid data cleaning.Thus,a lot of time used for data pre-process could be saved,and this would supply a stronger support on the further data processing and data interpretation.

关 键 词：物探数据 洁净数据 数据结构 PYTHON 数据清洗 

分 类 号：P631.4[天文地球—地质矿产勘探]