偏振多通道遥感云检测的阈值优化
|
摘要
大气中云的存在会严重影响气溶胶的反演精度。经验阈值法是一种常用的云检测方法,其较强的主观性和难以应对环境时空动态变化或星载探测仪差异的缺点,导致"云"和"晴"边缘分类误差增大,且检测自动化程度较低。针对下垫面为陆地的大气云检测,提出一种多通道偏振遥感图像的统计分类与数据融合的阈值优化方法,该方法首先通过半监督Kmeans聚类及其统计特征,决定像元属于"云"和"晴"两类的双亮度阈值;然后在阈值周边分类模糊区,用D-S证据理论获取多通道检测的联合置信度因子,求得模糊区像元分类的细化阈值;最终以顺序决策过程实现"云"和"晴"两类目标的精确分类。为了验证所提方法的有效性,利用POLDER3载荷遥感数据进行云检测实验,并与POLDER3产品结果进行比较。结果表明:所提方法与POLDER法的分类符合度为95%,目测发现这些误检大多发生在云边缘处,表明所提方法对云边缘处的分类具有较好的敏感性。
The existence of clouds in the atmosphere degrades the accuracy of aerosol retrieval.The empirical threshold method is popular in could detection,however its strong subjectivity and difficulty in coping with the dynamic spatial-temporal changes of the environment or the difference among satellite-borne detectors result in a large classification error at the boundary of‘cloud'and‘clear'.In addition,its automatic detection is also poor.To achieve an effective detection of cloud over the land surface in the atmosphere,we propose a threshold optimized method which combines the statistical classification with data fusion of polarized multichannel remote sensing images.As for this method,a dual-brightness threshold to distinguish‘cloud'from ‘clear'for most pixels is first derived based on the semi-supervised Kmeans clustering and its statistical features.Then,the joint confidence factor of multichannel data is calculated by the D-S evidence theory for each pixel in the fuzzy area of threshold neighborhood,and thus the fine threshold is acquired.The two objects of ‘cloud'and ‘clear'are finally and accurately classified in the sequential decision process.To validate the effectiveness of the proposed method,we perform a cloud detection experiment based on the remote sensing load data of POLRED3,and compare the measured results with the results of POLRED3.The results show that the classification by the proposed method is well consistent with that by the POLDER method with a high conformity of 95%.The error pixels are mostly located at the boundary between cloud and clear,indicating that the proposed method exhibits a favorable sensitivity to the classification at the cloud edge.
The existence of clouds in the atmosphere degrades the accuracy of aerosol retrieval.The empirical threshold method is popular in could detection,however its strong subjectivity and difficulty in coping with the dynamic spatial-temporal changes of the environment or the difference among satellite-borne detectors result in a large classification error at the boundary of‘cloud'and‘clear'.In addition,its automatic detection is also poor.To achieve an effective detection of cloud over the land surface in the atmosphere,we propose a threshold optimized method which combines the statistical classification with data fusion of polarized multichannel remote sensing images.As for this method,a dual-brightness threshold to distinguish‘cloud'from ‘clear'for most pixels is first derived based on the semi-supervised Kmeans clustering and its statistical features.Then,the joint confidence factor of multichannel data is calculated by the D-S evidence theory for each pixel in the fuzzy area of threshold neighborhood,and thus the fine threshold is acquired.The two objects of ‘cloud'and ‘clear'are finally and accurately classified in the sequential decision process.To validate the effectiveness of the proposed method,we perform a cloud detection experiment based on the remote sensing load data of POLRED3,and compare the measured results with the results of POLRED3.The results show that the classification by the proposed method is well consistent with that by the POLDER method with a high conformity of 95%.The error pixels are mostly located at the boundary between cloud and clear,indicating that the proposed method exhibits a favorable sensitivity to the classification at the cloud edge.
引文
[1] Guo H,Gu X F,Xie D H,et al.A review of atmospheric aerosol research by using polarization remotesensing[J]. SpectroscopyandSpectral Analysis,2014,34(7):1873-1880.郭红,顾行发,谢东海,等.大气气溶胶偏振遥感研究进展[J].光谱学与光谱分析,2014,34(7):1873-1880.
[2] Huang H L,Yi W N,Qiao Y L.Validation of retrieving aerosol optical parameters over the sea using airborne directional polarized camera[J].Acta Optica Sinica,2014,34(6):0601004.黄红莲,易维宁,乔延利.基于航空偏振相机的海上气溶胶光学特性反演与验证[J].光学学报,2014,34(6):0601004.
[3] Cheng T H,Chen L F,Gu X F,et al.Cloud phase classification and validation based on multi-angular polarized characteristics of cloud[J].Acta Optica Sinica,2008,28(10):1849-1855.程天海,陈良富,顾行发,等.基于多角度偏振特性的云相态识别及验证[J].光学学报,2008,28(10):1849-1855.
[4] Sun X,Zhao H J.Retrieval algorithm for optical parameters of aerosol over land surface from POLDER data[J].Acta Optica Sinica,2009,29(7):1772-1777.孙夏,赵慧洁.基于POLDER数据反演陆地上空气溶胶光学特性[J].光学学报,2009,29(7):1772-1777.
[5] Kriebel K T,Gesell G,Kstner M,et al.The cloud analysistoolAPOLLO:improvementsand validations[J].International Journal of Remote Sensing,2003,24(12):2389-2408.
[6] Buriez J C,Vanbauce C,Parol F,et al.Cloud detection and derivation of cloud properties from POLDER[J]. InternationalJournalofRemote Sensing,1997,18(13):2785-2813.
[7] Zhu Z,Woodcock C E.Object-based cloud and cloud shadow detection in Landsat imagery[J].Remote Sensing of Environment,2012,118:83-94.
[8] Shan N,Zheng T Y,Wang Z S.High-speed and high-accuracy algorithm for cloud detection and its application[J].Journal of Remote Sensing,2009,13(6):1138-1155.单娜,郑天垚,王贞松.快速高准确度云检测算法及其应用[J].遥感学报,2009,13(6):1138-1155.
[9] Xia Y,Cui S C,Yang S Z.Cloud detection method for high resolution satellite image based on multidimensional features[J].Journal of Atmospheric and Environmental Optics,2017,12(6):465-473.夏雨,崔生成,杨世植.综合高分卫星图像多维特征的云检测方法[J].大气与环境光学学报,2017,12(6):465-473.
[10] Wang W,Song W G,Liu S X,et al.A cloud detection algorithm for MODIS images combining Kmeans clustering and multi-spectral threshold method[J].Spectroscopy and Spectral Analysis,2011,31(4):1061-1064.王伟,宋卫国,刘士兴,等.Kmeans聚类与多光谱阈值相结合的MODIS云检测算法[J].光谱学与光谱分析,2011,31(4):1061-1064.
[11] Jin Z Q,Zhang L,Liu S C,et al.Cloud detection and cloud phase retrieval based on BP neural network[J].Optics&Optoelectronic Technology,2016,14(5):74-77.靳泽群,张玲,刘神聪,等.基于BP神经网络的云检测和云相态识别[J].光学与光电技术,2016,14(5):74-77.
[12] Xia M,Shen M Y,Wang J F,et al.Cloud fraction of satellite imagery based on convolutional neural networks[J].Journal of System Simulation,2018,30(5):1623-1630.夏旻,申茂阳,王舰锋,等.基于卷积神经网络的卫星云量计算[J].系统仿真学报,2018,30(5):1623-1630.
[13] Kang X G,Sun L X.Automatic cloud detection algorithm on artificial neural networks[J].Journal of PLA University of Science and Technology(Natural Science Edition),2005,6(5):506-510.康晓光,孙龙祥.基于人工神经网络的云自动检测算法[J].解放军理工大学学报(自然科学版),2005,6(5):506-510.
[14] Wang F C,Liu X T,Huang S C.Target fusion1228005-7detection with multi-feature based on fuzzy evidence theory[J].Acta Optica Sinica,2010,30(3):713-719.王凤朝,刘兴堂,黄树采.基于模糊证据理论的多特征目标融合检测算法[J].光学学报,2010,30(3):713-719.
[15] Liu H Y,Zhao Z G,Ba H X.Multisensor target identification method based on weighted evidence combination[J].Journal of PLA University of Science and Technology(Natural Science Edition),2005,6(6):521-524.刘海燕,赵宗贵,巴宏欣.一种基于加权证据合成的多传感器目标识别方法[J].解放军理工大学学报(自然科学版),2005,6(6):521-524.
[2] Huang H L,Yi W N,Qiao Y L.Validation of retrieving aerosol optical parameters over the sea using airborne directional polarized camera[J].Acta Optica Sinica,2014,34(6):0601004.黄红莲,易维宁,乔延利.基于航空偏振相机的海上气溶胶光学特性反演与验证[J].光学学报,2014,34(6):0601004.
[3] Cheng T H,Chen L F,Gu X F,et al.Cloud phase classification and validation based on multi-angular polarized characteristics of cloud[J].Acta Optica Sinica,2008,28(10):1849-1855.程天海,陈良富,顾行发,等.基于多角度偏振特性的云相态识别及验证[J].光学学报,2008,28(10):1849-1855.
[4] Sun X,Zhao H J.Retrieval algorithm for optical parameters of aerosol over land surface from POLDER data[J].Acta Optica Sinica,2009,29(7):1772-1777.孙夏,赵慧洁.基于POLDER数据反演陆地上空气溶胶光学特性[J].光学学报,2009,29(7):1772-1777.
[5] Kriebel K T,Gesell G,Kstner M,et al.The cloud analysistoolAPOLLO:improvementsand validations[J].International Journal of Remote Sensing,2003,24(12):2389-2408.
[6] Buriez J C,Vanbauce C,Parol F,et al.Cloud detection and derivation of cloud properties from POLDER[J]. InternationalJournalofRemote Sensing,1997,18(13):2785-2813.
[7] Zhu Z,Woodcock C E.Object-based cloud and cloud shadow detection in Landsat imagery[J].Remote Sensing of Environment,2012,118:83-94.
[8] Shan N,Zheng T Y,Wang Z S.High-speed and high-accuracy algorithm for cloud detection and its application[J].Journal of Remote Sensing,2009,13(6):1138-1155.单娜,郑天垚,王贞松.快速高准确度云检测算法及其应用[J].遥感学报,2009,13(6):1138-1155.
[9] Xia Y,Cui S C,Yang S Z.Cloud detection method for high resolution satellite image based on multidimensional features[J].Journal of Atmospheric and Environmental Optics,2017,12(6):465-473.夏雨,崔生成,杨世植.综合高分卫星图像多维特征的云检测方法[J].大气与环境光学学报,2017,12(6):465-473.
[10] Wang W,Song W G,Liu S X,et al.A cloud detection algorithm for MODIS images combining Kmeans clustering and multi-spectral threshold method[J].Spectroscopy and Spectral Analysis,2011,31(4):1061-1064.王伟,宋卫国,刘士兴,等.Kmeans聚类与多光谱阈值相结合的MODIS云检测算法[J].光谱学与光谱分析,2011,31(4):1061-1064.
[11] Jin Z Q,Zhang L,Liu S C,et al.Cloud detection and cloud phase retrieval based on BP neural network[J].Optics&Optoelectronic Technology,2016,14(5):74-77.靳泽群,张玲,刘神聪,等.基于BP神经网络的云检测和云相态识别[J].光学与光电技术,2016,14(5):74-77.
[12] Xia M,Shen M Y,Wang J F,et al.Cloud fraction of satellite imagery based on convolutional neural networks[J].Journal of System Simulation,2018,30(5):1623-1630.夏旻,申茂阳,王舰锋,等.基于卷积神经网络的卫星云量计算[J].系统仿真学报,2018,30(5):1623-1630.
[13] Kang X G,Sun L X.Automatic cloud detection algorithm on artificial neural networks[J].Journal of PLA University of Science and Technology(Natural Science Edition),2005,6(5):506-510.康晓光,孙龙祥.基于人工神经网络的云自动检测算法[J].解放军理工大学学报(自然科学版),2005,6(5):506-510.
[14] Wang F C,Liu X T,Huang S C.Target fusion1228005-7detection with multi-feature based on fuzzy evidence theory[J].Acta Optica Sinica,2010,30(3):713-719.王凤朝,刘兴堂,黄树采.基于模糊证据理论的多特征目标融合检测算法[J].光学学报,2010,30(3):713-719.
[15] Liu H Y,Zhao Z G,Ba H X.Multisensor target identification method based on weighted evidence combination[J].Journal of PLA University of Science and Technology(Natural Science Edition),2005,6(6):521-524.刘海燕,赵宗贵,巴宏欣.一种基于加权证据合成的多传感器目标识别方法[J].解放军理工大学学报(自然科学版),2005,6(6):521-524.