面向人工智能识别的岩浆岩断面形态特征研究  

Study on the sectional morphological characteristics of magmatic rocks for artificial intelligence identification

作　　者：张诗檬 韩征 刘钊 付博 王文文 ZHANG Shimeng;HAN Zheng;LIU Zhao;FU Bo;WANG Wenwen(Information Center of Beijing Institute of Geology,Beijing 100195,China)

机构地区：[1]北京市地质矿产勘查院信息中心,北京100195

出　　处：《城市地质》2022年第1期96-105,共10页Urban Geology

基　　金：北京市自然科学基金项目(8204076)资助。

摘　　要：为了研发手机识别岩石的软件,以岩浆岩基础数据库的建立为重点,开展岩浆岩断面外部特征的基础研究。根据可直接用于人工智能识别外部特征的标准,即能够直接被手机摄像头识别、计算机量化和获取特征的标准,筛选出6个具有人工识别意义的外部形态特征,分别是颜色特征、结晶程度、相对粒度、绝对粒度、构造特征和断口特征。根据每种特征不同的量化规则,针对已经筛选出的6种岩浆岩外部特征,将抽象的外部特征量化为具体的数值,形成每种特征的量化值区间,形成岩浆岩外部特征识别的分类体系,建立岩浆岩外部特征识别的基础特征库。In order to develop software for identifying rock by mobile phone,it is very important to carry out basic research on the external characteristics of magmatic cross section.Based on the following standards—direct recognition by mobile phone cameras,quantification and characterization by computers—to be used directly by AI to identify external features,six external morphological characteristics with artificial recognition significance are selected:color feature,crystal feature,grain size relative feature,grain size absolute feature,structure feature and fracture feature respectively.According to the different quantification rules of each feature,concrete values are obtained through quantification of the abstract external features for the six selected external features of magmatic rocks.Thus,a range of quantified values for each feature is formed,a classification system for the recognition of the external features is created,and the basic feature database of magmatic rock external feature recognition is established.

关 键 词：岩浆岩外部特征 岩浆岩断面特征分类体系 岩浆岩基础特征库 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] P588.1[自动化与计算机技术—控制科学与工程]