基于卷积神经网络的无人机遥感测绘图像特征提取方法  

Feature Extraction Method for Convolutional Neural Network UAV Remote Sensing Mapping Images

作　　者：马宏平 MA Hongping(Shandong Zhengwei Survey&Mapping Co.,Ltd.,Jinan 250000,China)

机构地区：[1]山东正维勘察测绘有限公司,济南250000

出　　处：《移动信息》2024年第6期257-259,共3页MOBILE INFORMATION

摘　　要：在地质勘探学领域,无人机遥感技术已成为获取地表数据的重要手段,而卷积神经网络(CNN)因其出色的图像处理能力被广泛应用于图像特征提取。文中介绍了一种基于卷积神经网络的无人机遥感测绘图像特征提取方法,分析了CNN的基本概念与结构、无人机遥感图像的特征类型、网络结构设计、数据预处理、特征提取及特征融合过程,旨在提高地质勘探中无人机遥感图像分析的自动化和准确性。In the field of geological prospecting,drone remote sensing technology has become an important means to obtain surface data,and convolutional neural networks(CNN)are widely used in image feature extraction due to their excellent image processing capabilities.This paper introduces a feature extraction method of drone remote sensing mapping images based on convolutional neural networks,and analyzes the basic concept and structure of CNN,feature types of drone remote sensing images,network structure design,data preprocessing,feature extraction and feature fusion process,aiming to improve the automation and accuracy of drone remote sensing image analysis in geological exploration.

关 键 词：卷积神经网络 无人机 遥感测绘图像特征提取 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]