无人机高分辨率遥感影像分类方法研究
|
摘要
无人机遥感技术的发展,能够为信息领域实时地提供极其丰富的空间信息,如何处理和应用无人机获得高分辨率影像是当今研究热点。面向对象的分类方法相比与传统基于像元的分类方法能使分类精度和信噪比得到显著改善。论文利用高分辨率无人机影像,分别以面向对象法和基于像元法分别影像进行信息提取,并以混淆矩阵和KAPPA系数来评价它们分类效果。证明了在高分辨率遥感影像分类中,面向对象的遥感影像分类方法与传统遥感影像分类方法相比能获得更好的分类效果。但是,建立更加完善的精度评价体系以及如何在不同的尺度和区域选取合适的分类方法,还要开展更多的研究和验证工作。
The development of remote sensing technology for drones can provide extremely rich spatial information for the information field in real time. How to process and apply high-resolution images obtained by drones is a hot research topic. The object-oriented classification method and the traditional pixel-based classification method can significantly improve the classificationaccuracy and signal-to-noise ratio. In this paper,high-resolution UAV images are used to extract information from object-orientedand pixel-based images respectively,and their classification effects are evaluated by confusion matrix and KAPPA coefficients. It isproved that in the high-resolution remote sensing image classification,the object-oriented remote sensing image classification method can obtain better classification effect than the traditional remote sensing image classification method. However,to establish amore complete accuracy evaluation system and how to select appropriate classification methods at different scales and regions,butalso more research and verification work need to be carried out.
The development of remote sensing technology for drones can provide extremely rich spatial information for the information field in real time. How to process and apply high-resolution images obtained by drones is a hot research topic. The object-oriented classification method and the traditional pixel-based classification method can significantly improve the classificationaccuracy and signal-to-noise ratio. In this paper,high-resolution UAV images are used to extract information from object-orientedand pixel-based images respectively,and their classification effects are evaluated by confusion matrix and KAPPA coefficients. It isproved that in the high-resolution remote sensing image classification,the object-oriented remote sensing image classification method can obtain better classification effect than the traditional remote sensing image classification method. However,to establish amore complete accuracy evaluation system and how to select appropriate classification methods at different scales and regions,butalso more research and verification work need to be carried out.
引文
[1]王浩宇.面向对象高分辨率遥感影像城区道路及车辆信息提取研究[D].北京:北京交通大学,2015.WANG Haoyu. Study on urban road and vehicle information extraction from high spatial resolution remote sensingimage based on object oriented:a case study[D]. Beijing:Beijing Jiaotong University,2015.
[2]杜培军,阿里木·赛买提.高分辨率遥感影像分类的多示例集成学习[J].遥感学报,2013(01):77-97.DU Peijun,A Limu-saimai-di. Multi-sample IntegratedLearning of High Resolution Remote Sensing Image Classification[J]. Journal of Remote Sensing, 2013(01):77-97.
[3]王惠英,苏伟,周军其.基于多分类器组合面向对象遥感影像信息提取方法[J].测绘工程,2009(05):22-26.WANG Huiying,SU Wei,ZHOU Junji. Object-orientedRemote Sensing Image Extraction Method Based onMulti-classifier[J]. Engineering of Surveying and Mapping,2009(05):22-26.
[4]张立斌,王文杰.一种面向机场旅客的个性化信息服务系统[J].科技创新与应用,2014(19):38-39.ZHANG Libin,WANG Wenjie. A personalized information service system for airport travelers[J]. Science andTechnology Innovation and Application,2014(19):38-39.
[5]于水,马范援.一种基于数据融合的医学图像分割方法[J].计算机辅助设计与图形学学报,2001(12):1073-1076.YU Shui,MA Fanyuan. A Medical Image SegmentationMethod Based on Data Fusion[J]. Journal of Computer-aided Design&Computer Graphics,2001(12):1073-1076.
[6]黄娟.基于GIS的机场信息服务系统[D].北京:北京交通大学,2011.HUANG Juan. Airport Information Service System Basedon GIS[D]. Beijing:Beijing Jiaotong University,2011.
[7]朱鹏,宫鹏.全球陆表湿地潜在分布区制图及遥感验证[J].中国科学:地球科学,2014(08):1610-1620.ZHU Peng,GONG Peng. Mapping and Remote SensingVerification of Potential Distribution of Global Land Surface Wetland[J]. Earth Science,2014(08):1610-1620.
[8]易俐娜.面向对象遥感影像分类不确定性分析[D].武汉:武汉大学,2011.YI Lina. Uncertainty Analysis of Object-oriented RemoteSensing Image Classification[D]. Wuhan:Wuhan University,2011.
[9]曹倩倩,黄袁升.遥感影像分类方法精度研究[J].牡丹江师范学院学报(自然科学版),2017(01):37-38.CAO Qianqian,HUANG Yuansheng. Research on Accuracy of Remote Sensing Image Classification Method[J].Journal of Mudanjiangs'college(Natural Science Edition),2017(01):37-38.
[10]EDIRIWICKREMA A,KHORRAM J,HIERARCHICALS. Hierarchical maximum likelihood classification for improved accuracies[J]. World Journal of Clinical Cases,2013(09):290-294.
[11]马铭,苟长龙.遥感数据最小距离分类的几种算法[J].测绘通报,2017(03):157-159.MA Ming,GOU Changlong. Several Algorithms for Minimum Distance Classification of Remote Sensing Data[J].Bulletin of Surveying and Mapping,2017(03):157-159.
[12]蒋宗礼.人工神经网络导论[M].北京:高等教育出版社,2008.JIANG Zongli. Introduction to Artificial Neural Networks[M]. Beijing:Higher Education Press,2008.
[13]SRINIVASU GANGI SETTI,R.N.RAO. Artificial neuralnetwork approach for prediction of stress-strain curve ofnear b Titanium alloy[J]. Rare Metals,2014(03):249-257.
[14]陈忠.高分辨率遥感图像分类技术研究[D].北京:中国科学院研究生院(遥感应用研究所),2006.CHEN Zhong. Research on High Resolution RemoteSensing Image Classification Technology[D]. Beijing:Graduate School of Chinese Academy of Sciences(Institute of Remote Sensing Applications),2006.
[15]MADRO?AL D,LAZCANO R,SALVADOR R,et al.SVM-based real-time hyperspectral image classifier ona manycore architecture[J]. Journal of Systems Architecture,2017,80:30-40.
[16]管珍,曹广超,易俊柱.面向对象的遥感影像分类研究[J].科技创新导报,2010(34):8-10,12.GUAN Zhen,CAO Anchao,YI Junzhu. Research on Object-oriented Remote Sensing Image Classification[J].Science and Technology Consulting Herald,2010(34):8-10,12.
[17]林喜兰.图像分割算法研究及其应用[D].无锡:江南大学,2016.LIN Xilan. Research on Image Segmentation Algorithmand Its Application[D]. Wuxi:Jiangnan University,2016.
[18]章毓晋.图像分割[J].计算机辅助设计与图形学学报,2015(03):379-387.ZHANG Yujin. Image segmentation[J]. Journal of Computer-aided Design&Computer Graphics,2015(03):379-387.
[2]杜培军,阿里木·赛买提.高分辨率遥感影像分类的多示例集成学习[J].遥感学报,2013(01):77-97.DU Peijun,A Limu-saimai-di. Multi-sample IntegratedLearning of High Resolution Remote Sensing Image Classification[J]. Journal of Remote Sensing, 2013(01):77-97.
[3]王惠英,苏伟,周军其.基于多分类器组合面向对象遥感影像信息提取方法[J].测绘工程,2009(05):22-26.WANG Huiying,SU Wei,ZHOU Junji. Object-orientedRemote Sensing Image Extraction Method Based onMulti-classifier[J]. Engineering of Surveying and Mapping,2009(05):22-26.
[4]张立斌,王文杰.一种面向机场旅客的个性化信息服务系统[J].科技创新与应用,2014(19):38-39.ZHANG Libin,WANG Wenjie. A personalized information service system for airport travelers[J]. Science andTechnology Innovation and Application,2014(19):38-39.
[5]于水,马范援.一种基于数据融合的医学图像分割方法[J].计算机辅助设计与图形学学报,2001(12):1073-1076.YU Shui,MA Fanyuan. A Medical Image SegmentationMethod Based on Data Fusion[J]. Journal of Computer-aided Design&Computer Graphics,2001(12):1073-1076.
[6]黄娟.基于GIS的机场信息服务系统[D].北京:北京交通大学,2011.HUANG Juan. Airport Information Service System Basedon GIS[D]. Beijing:Beijing Jiaotong University,2011.
[7]朱鹏,宫鹏.全球陆表湿地潜在分布区制图及遥感验证[J].中国科学:地球科学,2014(08):1610-1620.ZHU Peng,GONG Peng. Mapping and Remote SensingVerification of Potential Distribution of Global Land Surface Wetland[J]. Earth Science,2014(08):1610-1620.
[8]易俐娜.面向对象遥感影像分类不确定性分析[D].武汉:武汉大学,2011.YI Lina. Uncertainty Analysis of Object-oriented RemoteSensing Image Classification[D]. Wuhan:Wuhan University,2011.
[9]曹倩倩,黄袁升.遥感影像分类方法精度研究[J].牡丹江师范学院学报(自然科学版),2017(01):37-38.CAO Qianqian,HUANG Yuansheng. Research on Accuracy of Remote Sensing Image Classification Method[J].Journal of Mudanjiangs'college(Natural Science Edition),2017(01):37-38.
[10]EDIRIWICKREMA A,KHORRAM J,HIERARCHICALS. Hierarchical maximum likelihood classification for improved accuracies[J]. World Journal of Clinical Cases,2013(09):290-294.
[11]马铭,苟长龙.遥感数据最小距离分类的几种算法[J].测绘通报,2017(03):157-159.MA Ming,GOU Changlong. Several Algorithms for Minimum Distance Classification of Remote Sensing Data[J].Bulletin of Surveying and Mapping,2017(03):157-159.
[12]蒋宗礼.人工神经网络导论[M].北京:高等教育出版社,2008.JIANG Zongli. Introduction to Artificial Neural Networks[M]. Beijing:Higher Education Press,2008.
[13]SRINIVASU GANGI SETTI,R.N.RAO. Artificial neuralnetwork approach for prediction of stress-strain curve ofnear b Titanium alloy[J]. Rare Metals,2014(03):249-257.
[14]陈忠.高分辨率遥感图像分类技术研究[D].北京:中国科学院研究生院(遥感应用研究所),2006.CHEN Zhong. Research on High Resolution RemoteSensing Image Classification Technology[D]. Beijing:Graduate School of Chinese Academy of Sciences(Institute of Remote Sensing Applications),2006.
[15]MADRO?AL D,LAZCANO R,SALVADOR R,et al.SVM-based real-time hyperspectral image classifier ona manycore architecture[J]. Journal of Systems Architecture,2017,80:30-40.
[16]管珍,曹广超,易俊柱.面向对象的遥感影像分类研究[J].科技创新导报,2010(34):8-10,12.GUAN Zhen,CAO Anchao,YI Junzhu. Research on Object-oriented Remote Sensing Image Classification[J].Science and Technology Consulting Herald,2010(34):8-10,12.
[17]林喜兰.图像分割算法研究及其应用[D].无锡:江南大学,2016.LIN Xilan. Research on Image Segmentation Algorithmand Its Application[D]. Wuxi:Jiangnan University,2016.
[18]章毓晋.图像分割[J].计算机辅助设计与图形学学报,2015(03):379-387.ZHANG Yujin. Image segmentation[J]. Journal of Computer-aided Design&Computer Graphics,2015(03):379-387.