Classification of Lunar Lineament Structures Based on GoogLeNet
作者单位：Faculty of Information TechnologyMacau University of Science and Technology Lunar and Planetary Science Laboratory/Space Science InstituteMacau University of Science and Technology
会议日期：2019年
学科分类：07[理学] 070401[理学-天体物理] 0704[理学-天文学]
基 金：supported by Macao University of Science and Technology Development Fund of Macao (NO.099/2016/A3)
关 键 词：Convolutional Network GoogLeNet Lunar Lineament Structure
摘 要：This article aims to give a brief introduction on using GoogLeNet on Lunar lineament structure classification,mainly focusing on non-lineament structure,Lunar ridges and Lunar *** same network which was used to identify the non-lineament and lineament lunar structure has already reached the accuracy of around 84% based on a total sample of *** the so called three-type classification with around 300 unbalanced distribution sample,the result is around 85% to 88%,especially on identifying the lunar ridge,which has already reached over 98%.