基于异构机器学习算法融合的遥感影像分类
|
摘要
针对信息获取与处理过程中的不确定性导致的遥感数据分类精度难以满足土地覆盖变化、环境监测、专题信息提取等应用方面的需求,提出了一种基于机器学习的分类融合算法。采用6种异构分类器,以查准率及查全率矩阵为先验知识,依据分类器差异性指数AD对单分类器进行优化组合,结合三维概率矩阵分别得到抽象级、排序级和度量级的分类融合结果输出,并以北京地区Landsat 8遥感影像的典型区域为研究对象进行分类预测。结果表明,从6个单分类器中选取3个进行组合时的效果较好,其中AD值最大的(NB,KNN,SVM)分类器组合是综合效果最好的分类器组合;所提算法的抽象级输出比单分类的平均精度高12.28%,比分类效果最好的单分类器SVM高2.24%;所提算法对多个"强成员分类器"进行融合仍然能有效提高分类精度,比常用融合算法RF,Bagging和Boosting分别高出11.23%,7.56%和11.36%,对各种地物的分类精度有显著的提高。
In the application of multi-spectral remote sensing data,such as land cover change,environmental monitoring and thematic information extraction,the classification accuracy is not high enough due to the uncertainty of remote sen-sing information acquisition and processing.In order to further improve the classification accuracy,this paper proposed a fusion algorithm based on 6 heterogeneous machine learning classifiers.This algorithm provides classification results in abstract level,ranked level and measurement level by using prior knowledge set which is composed of precision and recall matrix,Accuracy and Difference(AD) index of the combination of classifiers,and the 3-dimensional probability matrix.Based on the Landsat 8 image data,the classification results in the study area of Beijing are forecasted by the proposed fusion algorithm and other different algorithms respectively.Experimental results shows that the 3-classifier combination composed of NB,KNN and SVM obtaines maximum AD value and the best classification effect.The abstract level output of the algorithm is 12.28% higher than the average accuracy of 6 single classifiers and even 2.24% higher than the best single classifier of SVM.Compared with the commonly used algorithms such as Random Forest(RF),Bagging and Boosting failed in the case of "strong member classifier",the proposed fusion algorithm performs still well with accuracy 11.23%,7.56% and 11.36% higher than RF,Bagging and Boosting respectively.The proposed fusion algorithm can effectively improve the classification accuracy of remote sensing data by making full use of the diversity of classifiers and prior knowledge such as precision and recall matrix in the process of classification.
In the application of multi-spectral remote sensing data,such as land cover change,environmental monitoring and thematic information extraction,the classification accuracy is not high enough due to the uncertainty of remote sen-sing information acquisition and processing.In order to further improve the classification accuracy,this paper proposed a fusion algorithm based on 6 heterogeneous machine learning classifiers.This algorithm provides classification results in abstract level,ranked level and measurement level by using prior knowledge set which is composed of precision and recall matrix,Accuracy and Difference(AD) index of the combination of classifiers,and the 3-dimensional probability matrix.Based on the Landsat 8 image data,the classification results in the study area of Beijing are forecasted by the proposed fusion algorithm and other different algorithms respectively.Experimental results shows that the 3-classifier combination composed of NB,KNN and SVM obtaines maximum AD value and the best classification effect.The abstract level output of the algorithm is 12.28% higher than the average accuracy of 6 single classifiers and even 2.24% higher than the best single classifier of SVM.Compared with the commonly used algorithms such as Random Forest(RF),Bagging and Boosting failed in the case of "strong member classifier",the proposed fusion algorithm performs still well with accuracy 11.23%,7.56% and 11.36% higher than RF,Bagging and Boosting respectively.The proposed fusion algorithm can effectively improve the classification accuracy of remote sensing data by making full use of the diversity of classifiers and prior knowledge such as precision and recall matrix in the process of classification.
引文
[1] HUABING H,YANLEI C,NICHOLAS C,et al.Mapping major land cover dynamics in Beijing using all Landsat imagesin Google Earth Engine[J].Remote Sensing of Environment,2017,202:166-176.
[2] ZHU L,MA L.Class centroid alignment based domain adaptation for classification of remote sensing images[J].Pattern Re-cognition Letters,2016,83:124-132.
[3] JASPER V V,ARNOLD K B,DANIEL G B,et al.A review of current calibration and validation practices in land-change mode-ling[J].Environmental Modelling & Software,2016,82:174-182.
[4] ZHAI J H,ZHAO W X.Soft Combination of Probabilistic Neural Network Classifiers for Face Recognition[J].Computer Scien-ce,2015,42(7):305-308.(in Chinese)翟俊海,赵文秀.软组合概率神经网络分类器人脸识别方法[J].计算机科学,2015,42(7):305-308.
[5] DANIEL S P,CéSAR F,MARíA JR,et al.Probabilistic class hierarchies for multiclass classification[J].Journal of Computational Science,2018,26:254-263.
[6] NIKOLAS M,FUILLERMO A N,ASAL B.A machine learning approach to energy pile design[J].Computers and Geotechnics,2018,97(2):189-203.
[7] WU X D,XIAO Q,WEN J G,et al.Advances in uncertainty analysis for the validation of remote sensing products:Take leaf area index for example[J].Journal of Remote Sensing,2014,18(5):1011-1023.(in Chinese)吴小丹,肖青,闻建光,等.遥感数据产品真实性检验不确定性分析研究进展[J].遥感学报,2014,18(5):1011-1023.
[8] HAO T,MATTHEW B,FENG Z,et al.Deriving and validating Leaf Area Index (LAI) at multiple spatial scales through lidar remote sensing:A case study in Sierra National Forest,CA[J].Remote Sensing of Environment,2014,143:131-141.
[9] MITCHELL B L,DAVID A K,STUART R P,et al.A comparison of resampling methods for remote sensing classification andaccuracy assessment [J].Remote Sensing of Environment,2018,208:145-153.
[10] MFAUVEL,J CHANUSSOT,J ABENEDIKTSSON.A spatial-spectral kernel-based approach for the classification of remote-sensing images[J].Pattern Recognition,2012,45:381-392.
[11] MELLOR A,BOUKIR S.Exploring diversity in ensemble classification:Applications in large area land cover mapping[J].ISPRS Journal of Photogrammetry and Remote Sensing,2017,129:151-161.
[12] TANG Y,LI X R.Set-based similarity learning in subspace for agricultural remote sensing classification[J].Neurocomputing,2016,173:332-328.
[13] ALIM S,CLAUDIO P,PAOLO G,et al.Supervised and Semi-Supervised Multi-View Canonical Correlation Analysis Ensemble for Heterogeneous Domain Adaptation in Remote Sensing Image Classification[J].Remote Sensing,2017,9(4):337.
[14] YANG Y J,ZHAN Y L,TIAN Q J,et al.Crop classification based on GF-1/WFV NDVI time series[J].Transactions of the Chinese Society of Agricultural Engineering,2015,31(24):155-161.(in Chinese)杨闫君,占玉林,田庆久,等.基于GF-1/WFV NDVI时间序列数据的作物分类[J].农业工程学报,2015,31(24):155-161.
[15] XU H Q.A Study on Information Extraction of Water Body with the Modified Normalized Difference Water Index(MNDWI)[J].Journal of Remote Sensing,2005,9(5):589-595.(in Chinese)徐涵秋.利用改进的归一化差异水体指数(MNDWI)提取水体信息的研究[J].遥感学报,2005,9(5):589-595.
[16] 周志华.机器学习[M].北京:清华大学出版社,2016:28-31.
[17] ZHANG Y.Study of Remote Sensing Image Classification Based on Machine Learning[D].Beijing:Beijing Forestry University,2014.(in Chinese)张雁.基于机器学习的遥感图像分类研究[D].北京:北京林业大学,2014.
[18] CHIRICI G,SCOTTI R,MONTAGHI A.Stochastic gradient boosting classification trees for forest fuel types mapping through airborne laser scanning and IRS LISS-III imagery[J].International Journal of Applied Earth Observation and Geoinformation,2013,25:87-97.
[19] DU P J,XIA J S,ZHANG W,et al.Multiple Classifier System For Remote Sensing Image Classification:A Review[J].Sensors,2012,12(4):4764-4792
[20] NITZE I,B BARRETT,CAWKWELL F.Temporal optimisation of image acquisition for land cover classification with Random Forest and MODIS time-series[J].International Journal of Applied Earth Observation and Geoinformation,2015,34:136-146.
[21] GU Y F,LIU H.Sample-screening MKL method via boosting strategy for hyperspectral image classification[J].Neurocomputing,2016,173:1630-1639.
[2] ZHU L,MA L.Class centroid alignment based domain adaptation for classification of remote sensing images[J].Pattern Re-cognition Letters,2016,83:124-132.
[3] JASPER V V,ARNOLD K B,DANIEL G B,et al.A review of current calibration and validation practices in land-change mode-ling[J].Environmental Modelling & Software,2016,82:174-182.
[4] ZHAI J H,ZHAO W X.Soft Combination of Probabilistic Neural Network Classifiers for Face Recognition[J].Computer Scien-ce,2015,42(7):305-308.(in Chinese)翟俊海,赵文秀.软组合概率神经网络分类器人脸识别方法[J].计算机科学,2015,42(7):305-308.
[5] DANIEL S P,CéSAR F,MARíA JR,et al.Probabilistic class hierarchies for multiclass classification[J].Journal of Computational Science,2018,26:254-263.
[6] NIKOLAS M,FUILLERMO A N,ASAL B.A machine learning approach to energy pile design[J].Computers and Geotechnics,2018,97(2):189-203.
[7] WU X D,XIAO Q,WEN J G,et al.Advances in uncertainty analysis for the validation of remote sensing products:Take leaf area index for example[J].Journal of Remote Sensing,2014,18(5):1011-1023.(in Chinese)吴小丹,肖青,闻建光,等.遥感数据产品真实性检验不确定性分析研究进展[J].遥感学报,2014,18(5):1011-1023.
[8] HAO T,MATTHEW B,FENG Z,et al.Deriving and validating Leaf Area Index (LAI) at multiple spatial scales through lidar remote sensing:A case study in Sierra National Forest,CA[J].Remote Sensing of Environment,2014,143:131-141.
[9] MITCHELL B L,DAVID A K,STUART R P,et al.A comparison of resampling methods for remote sensing classification andaccuracy assessment [J].Remote Sensing of Environment,2018,208:145-153.
[10] MFAUVEL,J CHANUSSOT,J ABENEDIKTSSON.A spatial-spectral kernel-based approach for the classification of remote-sensing images[J].Pattern Recognition,2012,45:381-392.
[11] MELLOR A,BOUKIR S.Exploring diversity in ensemble classification:Applications in large area land cover mapping[J].ISPRS Journal of Photogrammetry and Remote Sensing,2017,129:151-161.
[12] TANG Y,LI X R.Set-based similarity learning in subspace for agricultural remote sensing classification[J].Neurocomputing,2016,173:332-328.
[13] ALIM S,CLAUDIO P,PAOLO G,et al.Supervised and Semi-Supervised Multi-View Canonical Correlation Analysis Ensemble for Heterogeneous Domain Adaptation in Remote Sensing Image Classification[J].Remote Sensing,2017,9(4):337.
[14] YANG Y J,ZHAN Y L,TIAN Q J,et al.Crop classification based on GF-1/WFV NDVI time series[J].Transactions of the Chinese Society of Agricultural Engineering,2015,31(24):155-161.(in Chinese)杨闫君,占玉林,田庆久,等.基于GF-1/WFV NDVI时间序列数据的作物分类[J].农业工程学报,2015,31(24):155-161.
[15] XU H Q.A Study on Information Extraction of Water Body with the Modified Normalized Difference Water Index(MNDWI)[J].Journal of Remote Sensing,2005,9(5):589-595.(in Chinese)徐涵秋.利用改进的归一化差异水体指数(MNDWI)提取水体信息的研究[J].遥感学报,2005,9(5):589-595.
[16] 周志华.机器学习[M].北京:清华大学出版社,2016:28-31.
[17] ZHANG Y.Study of Remote Sensing Image Classification Based on Machine Learning[D].Beijing:Beijing Forestry University,2014.(in Chinese)张雁.基于机器学习的遥感图像分类研究[D].北京:北京林业大学,2014.
[18] CHIRICI G,SCOTTI R,MONTAGHI A.Stochastic gradient boosting classification trees for forest fuel types mapping through airborne laser scanning and IRS LISS-III imagery[J].International Journal of Applied Earth Observation and Geoinformation,2013,25:87-97.
[19] DU P J,XIA J S,ZHANG W,et al.Multiple Classifier System For Remote Sensing Image Classification:A Review[J].Sensors,2012,12(4):4764-4792
[20] NITZE I,B BARRETT,CAWKWELL F.Temporal optimisation of image acquisition for land cover classification with Random Forest and MODIS time-series[J].International Journal of Applied Earth Observation and Geoinformation,2015,34:136-146.
[21] GU Y F,LIU H.Sample-screening MKL method via boosting strategy for hyperspectral image classification[J].Neurocomputing,2016,173:1630-1639.