选取平顶山平煤矿区附近区域作为研究区,使用无人机进行拍摄,以获悉矿区周边地物种类,补充辅助决策中的居民地、道路等信息,重点研究遥感影像的分析处理。为实现遥感影像分类并增强可视化效果,以无人机影像为源图像,分别采用无人机正射影像的主流处理软件Photoscan和Pix4d,完成图像的预处理及拼接,发现利用Pix4d软件进行图像拼接效果更佳。采用非监督分类及监督分类中的最大似然法和支持向量机,研究影像拼接分类方法,并结合ArcGIS软件,增强可视化效果。通过分析结果图像的类内精度、总体精度、kappa系数指标,完成分类质量评价,发现支持向量机分类效果更佳。此次针对平顶山地区遥感影像的试验结果,对于完善灾害预评估起到一定借鉴作用,并可为震后应急救灾辅助决策提供有效的数据支撑。
This paper selects the vicinity of the Pingmei mining area as the research area. Images of the research area are obtained by a UAV to learn the land types and supplement the information of residential areas and roads in auxiliary decision-making. The analysis and processing of remote sensing images are mainly studied. In order to realize the classification of remote sensing images and enhance the visualization effect, the UAV images are taken as the source images and the mainstream processing software Photoscan and Pix4d are used to complete the image processing and mosaic. It is found that the image mosaic effect is better by Pix4d software. The maximum likelihood method and support vector machine in unsupervised classification and supervised classification are used to study the image mosaic classification method, and ArcGIS software is used to enhance the visualization effect. Through analyzing the intra-class precision, overall precision, and kappa coefficient index of the result images, the classification quality evaluation is completed, and it is found that the classification effect of the support vector machine is better. The experimental results of remote sensing images in the Pingdingshan area can be used for reference to improve disaster pre-assessment and provide effective data support for post-earthquake emergency decision-making.
2021,42(2): 72-80 收稿日期:2020-11-02
DOI:10.3969/j.issn.1003-3246.2021.02.008
基金项目:中国地震局地震应急青年重点任务(项目编号:CEAEDEM202013);2020年度河南省青年人才托举工程项目(项目编号:2020HYTP036)
作者简介:李晓阳(1989-04-),男,硕士,主要从事应急技术运维工作。E-mail:469680140@qq.com
*通讯作者:樊华(1988-04-),女,硕士,主要从事应急技术运维工作。E-mail:867877458@qq.com
参考文献:
何家乐. 基于无人机遥感影像的地震形变场三维测量方法[J]. 华南地震,2019,39(4):34-39.
黄海峰,易武,张国栋,等. 引入小型无人机遥感的滑坡应急治理勘查设计方法[J]. 防灾减灾工程学报,2017,39(1):99-104.
陆晓果,王同科,梁社芳,等. 最大似然分类的训练样本敏感度研究[J]. 中国农业信息,2018,30(2):76-85.
吴俣,余涛,谢东海. 面向应急响应的无人机图像快速自动拼接[J]. 计算机辅助设计与图形学学报,2013,25(3):410-416.
许军强,张斌,袁晶,等. 基于遥感技术的南水北调水源区(河南段)石漠化遥感调查与评价[J]. 世界地质,2019,38(4):1 091-1 098.
薛腾飞,张景发,李强. 基于相关变化检测与面向对象分类技术的多源遥感图像震害信息提取[J]. 地震学报,2016,38(3):496-505.
张合兵,张克,刘培,等. 基于RS和GIS的矿区生态指标提取与安全评价——以焦作矿区为例[J]. 煤炭科学技术,2020,48(4):80-88.
张顺,赵玉金,白永飞,等. 基于低空无人机的草原灌丛遥感辨识方法[J]. 热带地理,2019,39(4):512-520.
张晓羽,李凤日,甄贞,等. 基于随机森林模型的陆地卫星-8遥感影像森林植被分类[J]. 东北林业大学学报,2016,44(6):53-57.
张雪华,王晓青,杜晓霞,等. 基于无人机遥感影像及其点云特征的建筑物震害提取[J]. 地震研究,2019,26(2):230-235.
Ball G, Hall D. ISODATA, a novel method of data analysis and pattern classification[R]. Stanford University, 1965.