高分辨率全极化合成孔径雷达数据海冰二次分类方法研究

随着高分辨率全极化合成孔径雷达(synthetic aperture radar, SAR)影像的大量出现, 传统分类方法因所得分类结果中包含斑点噪声, 导致分类精度大大降低, 因而利用基于伴生关系的二次分类方法对辽东湾RADARSAT-2卫星的海冰SAR影像进行分类, 并将所得分类结果与专家解译分类结果和经典的基于H-α分解的Wishart监督分类结果进行对比, 证明本方法可以消除斑点噪声的影响, 同时具有较好的视觉效果和较高的分类精度。     

渤海海冰微波散射特性研究

Microwave remote sensing has become the primary means for sea-ice research, and has been supported by a great deal of field experiments and theoretical studies regarding sea-ice microwave scattering. However, these studies have been barely carried in the Bohai Sea. The sea-ice microwave scattering mechanism was first developed for the thin sea ice with slight roughness in the Bohai Sea in the winter of 2012, and included the backscattering coefficients which were measured on the different conditions of three bands(L, C and X), two polarizations(HH and VV), and incident angles of 20° to 60°, using a ground-based scatterometer and the synchronous physical parameters of the sea-ice temperature, density, thickness, salinity, and so on. The theoretical model of the sea-ice electromagnetic scattering is obtained based on these physical parameters. The research regarding the sea-ice microwave scattering mechanism is carried out through two means, which includes the comparison between the field microwave scattering data and the simulation results of the theoretical model, as well as the feature analysis of the four components of the sea-ice electromagnetic scattering. It is revealed that the sea-ice microwave scattering data and the theoretical simulation results vary in the same trend with the incident angles. Also, there is a visible variant in the sensitivity of every component to the different bands.For example, the C and X bands are sensitive to the top surface, the X band is sensitive to the scatterers, and the L and C bands are sensitive to the bottom surface, and so on. It is suggested that the features of the sea-ice surfaces and scatterers can be retrieved by the further research in the future. This experiment can provide an experimental and theoretical foundation for research regarding the sea-ice microwave scattering characteristics in the Bohai Sea.     

结合纹理特征的SVM海冰分类方法研究

海冰分类是遥感监测领域中的重要应用之一,海冰分类的准确性对于评估海冰冰情、保证航海安全和开辟北极航道具有重要的意义。针对海冰分类问题,本文选用Sentinel-1遥感数据,结合纹理特征分析,提出了一种改进的SAR海冰分类方法。该方法选用灰度共生矩阵提取特征值,通过实验得到适宜用于海冰分类的多特征组合,在此基础上利用支持向量机开展SAR海冰类型的分类研究。实验结果表明,该方法可以实现对海冰SAR图像中一年冰、多年冰和海水3种类型识别,与传统的海冰分类方法神经网络和最大似然法相比较,使用SVM分类方法,结合纹理特征开展海冰类型监测是可行的,同时也表明多特征组合有利于提高SAR图像的分类精度,从而验证了本方法的有效性,为海冰分类提供了一种新思路。     

基于简缩极化SAR的溢油检测与分类方法

针对简缩极化SAR在海上溢油的检测与分类应用开展研究,利用欧式距离全面分析了简缩极化SAR的36种极化特征在溢油检测与油膜分类中的性能,发现简缩极化特征中的奇次散射系数的溢油检测性能最好,简缩极化熵的疑似溢油鉴别性能最好.在此基础上,提出了结合二叉树原理的简缩极化SAR溢油检测与油膜分类算法,并分析了RADARSAT-...     

C波段紧缩极化SAR海冰探测能力评估

The C-band synthetic aperture radar(SAR) data from the Bohai Sea of China, the Labrador Sea in the Arctic and the Weddell Sea in the Antarctic are used to analyze and discuss the sea ice full polarimetric information reconstruction ability under compact polarimetric modes. The type of compact polarimetric mode which has the highest reconstructed accuracy is analyzed, along with the performance impact of the reconstructed pseudo quad-pol SAR data on the sea ice detection and sea ice classification. According to the assessment and analysis, it is recommended to adopt the CTLR mode for reconstructing the polarimetric parameters σ_(HH)~0,σ_(VV)~0, H and α,while for reconstructing the polarimetric parameters σ_(HV)~0, ρ_(H-V), λ_1 and λ_2, it is recommended to use the π/4 mode.Moreover, it is recommended to use the π/4 mode in studying the action effects between the electromagnetic waves and sea ice, but it is recommended to use the CTLR mode for studying the sea ice classification.     

全极化SAR图像中溢油极化特征研究

相比于单极化SAR图像,全极化SAR图像不仅能体现海面目标的几何特征、后向散射特征,还能体现目标的极化特征。因此,在溢油检测方面,极化SAR更具优势。特征提取作为溢油检测的关键步骤,直接影响到溢油检测的精度。在本文中,我们分析了全极化SAR图像中海面溢油的极化特征,如极化散射熵、平均散射角等。并提出了新的极化特征P,该特征参数能够反映海面目标电磁散射过程中布拉格散射机制和镜面散射机制的比例。为了研究极化特征溢油检测的能力,本文基于SIR-C/X-SAR和Radarsat-2全极化SAR图像开展了相关实验,并对比分析了溢油的多种极化特征。实验结果显示,在中低风速情况下,C波段溢油探测效果优于L波段;本文提出的极化特征P对海面散射机制敏感;基准高度和特征参数P在C波段比其他极化特征更适于溢油检测。     

渤海海冰特征厚度分析

通过海冰生消机理和数值试验,讨论了渤海海冰特征厚度的存在条件;对不同厚度的海冰表面温度、冰面热量收支、冰面下热传导和太阳辐射透射量进行了对比分析,分析了渤海海冰向特征冰厚的动态演化过程;在不同气温、风速、相对湿度和海洋热通量等气象和水文条件下,对渤海特征冰厚进行了计算;讨论了海冰生消的动态平衡过程,分析了1997/1998年冬季辽东湾JZ20-2海域实测冰厚与特征冰厚的相互关系。对渤海特征冰厚分析将有助于渤海海冰数值模拟工作的开展和对不同重现期设计冰厚的推算。     

基于HY-2B波形特征的北极海冰分类算法

海冰类型识别对进行全球气候研究至关重要,利用高度计进行北极海冰监测是目前研究的热点。本文通过利用2019年12月和2020年3月的HY-2B高度计数据,探索了国产卫星高度计在海冰类型识别中的可用性,提取了HY-2B的波形功率最大值、脉冲峰值(Pulse Peakiness,PP)、前缘宽度(Leading Edge Width,LEW)和后向散射系数(Sigma0)4个波形特征,分析了HY-2B卫星高度计精确识别薄一年冰(Thin First-year ice,TFYI)、一年冰(First-year ice,FYI)、多年冰(Multi-year ice,MYI)、冰间水道(LEAD)和开阔水域(Open Water,OW)的能力。通过与俄罗斯北极和南极研究所(Arctic and Antarctic Research Institute,AARI)冰况图产品和MODIS冰间水道产品对比发现,综合PP、LEW及Sigma0和K最近邻法(K-Nearest Neighbor,KNN),平均最高海冰分类精度可达到91.96%。     

雷达观测海冰分类的研究

本文在国内首次应用陆基雷达对辽东湾海冰进行分类研究,认为冰表面粗糙度、冰厚以及海况是影响海冰散射的主要因素.研究结果以海冰雷达散射系数为依据,对海冰分类准确率达80%以上,并首次建立了海冰特征数据库.     

灰度共生矩阵纹理特征对SAR海冰漂移监测的增强性能研究

海冰漂移监测对气候变化分析、船只航行、海上石油平台等海上活动安全作业具有重要意义。当前主流的SAR海冰漂移监测方法多是基于SAR灰度图开展的,其受噪声、环境等因素的影响较大,导致其在海冰漂移探测时,特征失配率高,匹配正确率低。针对这一问题,本文尝试利用SAR海冰纹理特征来增强海冰漂移探测性能。首先对比分析了8种纹理特征对海冰漂移探测中特征匹配的增强性能,筛选出能够有效增强特征匹配性能的最优纹理特征;其次进一步分析了海冰类型、入射角和分辨率对基于纹理特征的海冰漂移探测性能增强的影响。实验结果表明,均值是最优的纹理特征,与SAR强度图相比,特征匹配正确率提高了约7%。     

Monitoring regional sea ice of the Bohai Sea by SSM/I scattering index

SSM/I data on the Bohai Sea of China from 1994 to 1996 have been studied.The sea ice algorithms of calibration/validation such as the brightness temperature index [AES/YORK TBI (Hollinger et al.,1992)] and ocean scattering index [OSI (Ferraro et al.,1996, The Bulletin of the American Meteorological Society, 77(5), 891-905)] are not applicable to the detection of regional sea ice of mid-latitude Bohai area.This paper presents a method of the scatfeting index SI and polarization index PI to identify regional Bohai Sea ice.Numerical results of vector radiative transfer for a model of a layer of sea ice over ocean simulate the relationship between SI/PI and sea ice depth, and others.It is employed to categorize the SI/PI indexes for the detection of sea ice.     

Monitoring regional sea ice of the Bohai Sea by SSM/I scattering index

ＭｏｎｉｔｏｒｉｎｇｒｅｇｉｏｎａｌｓｅａｉｃｅｏｆｔｈｅＢｏｈａｉＳｅａｂｙＳＳＭ／ＩｓｃａｔｔｅｒｉｎｇｉｎｄｅｘＪｉｎＹａｑｉｕ（ＲｅｃｅｉｖｅｄＪａｎｕａｒｙ６,１９９７;ａｃｃｅｐｔｅｄＭａｒｃｈ３０,１９９７）Ａｂｓｔｒａｃｔ──ＳＳＭ／...     

基于MODIS 影像渤海2012 年海冰变化分析

海冰是影响渤海冬季海运的严重灾害,研究年度周期内海冰时空变化具有重要意义。依据研究内容选择11期具有代表性的中分辨率成像光谱仪(MODerate-resolution Imaging Spectroradiometer,MODIS)影像数据,采用ENVI(The Environment for Visualizing Images)软件进行预处理,利用目视解译方法提取海冰面积与空间位置信息,采用ArcGIS软件制作了海冰时空变化图、海冰结冻与融化过程变化图。最后,对海冰面积变化、海冰结冻与融化过程进行了详细分析。研究结果表明,利用MODIS影像进行年度周期内海冰变化分析技术方法是可行的。     

一种新的用于极化SAR图像船只检测的散射相似性测度

本文提出了一种新的用于极化合成孔径雷达（SAR）海上船只目标检测的测度。首先利用散射相似性参数研究船只与海杂波散射机制的差异。基于这些差异,提出了一种新的船只检测测度,该测度能够有效区分船只目标和海杂波。然后提出了利用核密度估计（KDE）方法对该测度进行建模的方法。基于统计模型,实现了自适应恒虚警率（CFAR）的检测方案。本文利用多景C波段RADARSAT-2极化SAR（Pol-SAR）数据上进行实验,系统分析了所提出测度的海杂波拟合性能与船只检测性能,并与两个经典的极化测度进行了比较,实验和比较结果证明了所提的测度的可行性。     

卷积神经网络在卫星遥感海冰图像分类中的应用探究

本文以TensorFlow为框架搭建卷积神经网络,基于迁移学习的思想,以经典的手写数字识别作为引入,对不同代价函数和激活函数组合对卷积神经网络模型分类结果影响进行了评价分析。以HJ-1A/B渤海海冰图像为实验数据源,分析了不同函数组合对遥感海冰图像分类的影响,优选出交叉熵代价函数与ReLU激活函数为最佳的组合,证明了卷积神经网络在遥感海冰分类中的应用可行性。对渤海海冰图像分类结果进行验证,其中带标签样本验证精度为98.4%。使用该模型对无标签的测试样本进行识别,讨论了样本的窗口尺寸对海冰分类结果的影响,发现在400×400小范围分类实验中最佳窗口尺寸为2×2;最后对整个渤海海域进行识别验证,效果较好。     

海冰动力过程的改进离散元模型及在渤海的应用

海冰的断裂、重叠和堆积等离散分布特性广泛地存在于极区和副极区的不同海域,并对海冰的生消、运移过程有着重要影响。针对海冰在不同尺度下的离散分布特点,发展海冰动力过程的离散元方法有助于完善海冰数值模式,提高海冰数值模拟的计算精度。为此,本文针对海冰生消运移过程中的非连续分布和形变特性,发展了适用于海冰动力过程的改进离散元模型(MDEM)。不同于传统离散元方法,该模型将海冰离散为具有一定厚度、尺寸和密集度的圆盘单元。海冰单元设为诸多浮冰块的集合体,其在运移和相互接触碰撞过程中,依照质量守恒发生单元尺寸、密集度和厚度的相应变化。基于海冰离散性和流变性的特点,该模型采用黏弹性接触本构模型计算单元间的作用力,并依据Mohr-Coulomb准则计算海冰法向作用下的塑性变形及切向摩擦力。为验证该模型的可靠性,本文对海冰在规则水域内的运移和堆积过程进行了分析,离散元计算结果与解析值相一致;此外,对旋转风场下海冰漂移规律的模拟进一步验证了本文方法的精确性。在此基础上,对渤海辽东湾的海冰动力过程进行了48h数值分析,计算结果与卫星遥感资料和油气作业区的海冰现场监测数据吻合良好。在下一步工作中将考虑海冰离散元模拟中的热力因素影响,发展具有冻结、断裂效应的海冰离散元模型,更精确地模拟海冰动力-热力耦合作用下的生消和运移过程。     

渤海模型冰运动特征的现场观测研究

Sea ice drift is mainly controlled by ocean currents, local wind, and internal ice stress. Information on sea ice motion, especially in situ synchronous observation of an ice velocity, a current velocity, and a wind speed, is of great significance to identify ice drift characteristics. A sea ice substitute, the so-called "modelled ice", which is made by polypropylene material with a density similar to Bohai Sea ice, is used to complete a free drift experiment in the open sea. The trajectories of isolated modelled ice, currents and wind in the Bohai Sea during non-frozen and frozen periods are obtained. The results show that the currents play a major role while the wind plays a minor role in the free drift of isolated modelled ice when the wind is mild in the Bohai Sea. The modelled ice drift is significantly affected by the ocean current and wind based on the ice–current–wind relationship established by a multiple linear regression. The modelled ice velocity calculated by the multiple linear regression is close to that of the in situ observation, the magnitude of the error between the calculated and observed ice velocities is less than12.05%, and the velocity direction error is less than 6.21°. Thus, the ice velocity can be estimated based on the observed current velocity and wind speed when the in situ observed ice velocity is missing. And the modelled ice of same thickness with a smaller density is more sensitive to the current velocity and the wind speed changes. In addition, the modelled ice drift characteristics are shown to be close to those of the real sea ice, which indicates that the modelled ice can be used as a good substitute of real ice for in situ observation of the free ice drift in the open sea, which helps solve time availability, safety and logistics problems related to in situ observation on real ice.     

基于岸基雷达数据的鲅鱼圈海域冰情基本特征

以2010~2015年的岸基雷达海冰监测数据为基础,简要分析了鲅鱼圈海域的冰期、冰型和冰厚等冰情基本特征。近年来的海冰监测数据表明,鲅鱼圈海域的冰期明显偏短,冰期内不同阶段的冰型变化较明显,鲅鱼圈海域部分时段海冰分布比例变化较大,这主要归因于该区域的海冰运动。针对目前岸基雷达海冰监测的技术发展现状,提出了岸基雷达海冰监测技术研发的主要问题是雷达像元回波值不稳定与海冰实测样本量匮乏。     

The estimate of sea ice resources quantity in the Bohai Sea based on NOAA/AVHRR data

The research on sea ice resources is the academic base of sea ice exploitation in the Bohai Sea. According to the ice-water spectrum differences and the correlation between ice thickness and albedo, this paper comes up with a sea ice thickness inversion model based on the NOAA/AVHRR data. And then a sea ice resources quantity (SIQ) time series of Bohai Sea is established from 1987 to 2009. The results indicate that the average error of inversion sea ice thickness is below 30%. The maximum sea ice resources quantity is about 6 × 10 9 m 3 and the minimum is 1.3 × 10 9 m 3 . And a preliminary analysis has been made on the errors of the estimate of sea ice resources quantity (SIQ).