基于无人机影像和分类器集成的土地使用类型自动识别研究  

Research on Automatic Recognition of Land Use Types Based on the Integration of Drone Images and Classifiers

作　　者：周丹 ZHOU Dan(Guizhou Province Nonferrous Metals and Nuclear Industry Geological Exploration Bureau Fifth Brigade,Anshun 561000,China)

机构地区：[1]贵州省有色金属和核工业地质勘查局五总队,贵州安顺561000

出　　处：《城市勘测》2025年第1期33-37,共5页Urban Geotechnical Investigation & Surveying

摘　　要：为提升土地使用类型自动识别性能,研究基于无人机影像和分类器集成的土地使用类型自动识别方法,使土地使用类型识别精确度更高。利用SLA-4800型号无人机采集城市土地区域无人机影像,以土地全区域涵盖以及乡镇集合等准则为基础,构建土地使用类型自动识别标准,对无人机采集的土地无人机影像进行零均值化以及影像增强处理,实现土地无人机影像预处理,通过采用自动编码机与深度空谱特征联合算法,对预处理后的土地无人机影像进行特征提取,构建以SVM分类器、C4.5决策树分类器以及深度学习分类器组成的集成分类器,最终利用投票辨识实现土地使用类型自动识别。实验表明:该方法处理后土地遥感图像更加清晰,可实现土地使用类型自动识别,且具备较高的识别Kappa系数,有效助力城市各行业的土地配置。To improve the automatic recognition performance of land use types,a land use type automatic recognition method based on the integration of drone images and classifiers is studied to improve the accuracy of land use type recognition.Using the SLA-4800 model drone to collect drone images of urban land areas,based on the criteria of covering the entire land area and gathering townships,a standard for automatic recognition of land use types is constructed.The drone images collected by the drone are subjected to zero mean and image enhancement processing to achieve land drone image preprocessing.By using a joint algorithm of automatic encoding machines and deep spatial spectral features,extract features from preprocessed land drone images,construct an ensemble classifier consisting of SVM classifier,C4.5 decision tree classifier,and deep learning classifier,and finally use voting recognition to achieve automatic recognition of land use types.The experiment shows that the method can process clearer land remote sensing images,achieve automatic recognition of land use types,and have a high recognition Kappa coefficient,effectively assisting in land allocation in various industries of the city.

关 键 词：无人机影像 分类器集成 土地使用类型 自动识别 零均值化 SVM分类器 

分 类 号：TP753[自动化与计算机技术—检测技术与自动化装置] TP181[自动化与计算机技术—控制科学与工程]