In this paper, the long-term variation trend of the Antarctic sea ice in 1973～1994 and the inter-decade variation rule of the global sea level are analyzed. It is foundthat, the Antarctic sea ice area in 1980’s was ...In this paper, the long-term variation trend of the Antarctic sea ice in 1973～1994 and the inter-decade variation rule of the global sea level are analyzed. It is foundthat, the Antarctic sea ice area in 1980’s was significantly less than in 1970’s but with regional difference:decreasing in Regions Ⅰ, Ⅱ, and Ⅲ, and increasing in Region Ⅳ, the average global sea levelheightvalue in 1980’s was also significantly higher than in 1970’s but also with very large regional difference.Connecting variation of both to analyze their physical mechanism, it is pointed that, the accumulated seaice anomaly value in 1980’s less than in 1970’s means a global climate warming, the sea watertemperature and air temperature rising, sea water volume expanding, and more icebergs transportingfrom the ice cover on the Antarctic continent to ocean in the warmer years. As a result, the global sealevel raised significantly with the global average sea level value in 1980’s was 22 mm higher than in 1970’s. The Sea Level Raising (SLR) distributed uneven. It is especially true in the Pacific Ocean andAtlantic. This kind of uneven SLR distribution closely relates to, or is introduced by the uneven sea ice change of the Antarctic Regions.展开更多
高光谱异常变化检测能够从多时相高光谱遥感图像中寻找到数量稀少、与整体背景变化趋势不同、难以发现且令人感兴趣的异常变化。数据集规模较小、存在噪声干扰以及线性预测模型存在局限性等问题,极大地降低了传统高光谱异常变化检测方...高光谱异常变化检测能够从多时相高光谱遥感图像中寻找到数量稀少、与整体背景变化趋势不同、难以发现且令人感兴趣的异常变化。数据集规模较小、存在噪声干扰以及线性预测模型存在局限性等问题,极大地降低了传统高光谱异常变化检测方法的检测性能。目前,自编码器已被成功地应用于高光谱异常变化检测。然而,单个自编码器在处理多时相高光谱图像时,仅关注图像的重构质量,在获取瓶颈特征时往往忽略了图像中复杂的光谱变化信息。为了解决该问题,提出了一种基于双空间共轭自编码器的多时相高光谱异常变化检测(Multi-temporal Hyperspectral Anomaly Change Detection Based on Dual Space Conjugate Autoencoder,DSCAE)方法。所提方法包含两个共轭的自编码器,即它们从不同方向构造各自的潜在特征。在该方法的训练过程中,首先,两幅不同时刻的高光谱图像经过各自的编码器分别获得相应的潜在空间特征表示,再分别经过各自的解码器获得另一时刻的预测图像;其次,在样本空间和潜在空间中施加不同的约束条件,并在两个空间中最小化相应的损失函数;最后,两幅输入图像经过共轭自编码器后获得各自的异常损失图,对所得的两幅异常损失图采用取小运算得到最终的异常变化强度图,以便在减小输入图像间背景光谱差异的同时突出异常变化。在高光谱异常变化检测基准数据集上的实验结果表明,与10种相关方法相比,DSCAE展现了更优的检测性能。展开更多
文摘In this paper, the long-term variation trend of the Antarctic sea ice in 1973～1994 and the inter-decade variation rule of the global sea level are analyzed. It is foundthat, the Antarctic sea ice area in 1980’s was significantly less than in 1970’s but with regional difference:decreasing in Regions Ⅰ, Ⅱ, and Ⅲ, and increasing in Region Ⅳ, the average global sea levelheightvalue in 1980’s was also significantly higher than in 1970’s but also with very large regional difference.Connecting variation of both to analyze their physical mechanism, it is pointed that, the accumulated seaice anomaly value in 1980’s less than in 1970’s means a global climate warming, the sea watertemperature and air temperature rising, sea water volume expanding, and more icebergs transportingfrom the ice cover on the Antarctic continent to ocean in the warmer years. As a result, the global sealevel raised significantly with the global average sea level value in 1980’s was 22 mm higher than in 1970’s. The Sea Level Raising (SLR) distributed uneven. It is especially true in the Pacific Ocean andAtlantic. This kind of uneven SLR distribution closely relates to, or is introduced by the uneven sea ice change of the Antarctic Regions.
文摘高光谱异常变化检测能够从多时相高光谱遥感图像中寻找到数量稀少、与整体背景变化趋势不同、难以发现且令人感兴趣的异常变化。数据集规模较小、存在噪声干扰以及线性预测模型存在局限性等问题,极大地降低了传统高光谱异常变化检测方法的检测性能。目前,自编码器已被成功地应用于高光谱异常变化检测。然而,单个自编码器在处理多时相高光谱图像时,仅关注图像的重构质量,在获取瓶颈特征时往往忽略了图像中复杂的光谱变化信息。为了解决该问题,提出了一种基于双空间共轭自编码器的多时相高光谱异常变化检测(Multi-temporal Hyperspectral Anomaly Change Detection Based on Dual Space Conjugate Autoencoder,DSCAE)方法。所提方法包含两个共轭的自编码器,即它们从不同方向构造各自的潜在特征。在该方法的训练过程中,首先,两幅不同时刻的高光谱图像经过各自的编码器分别获得相应的潜在空间特征表示,再分别经过各自的解码器获得另一时刻的预测图像;其次,在样本空间和潜在空间中施加不同的约束条件,并在两个空间中最小化相应的损失函数;最后,两幅输入图像经过共轭自编码器后获得各自的异常损失图,对所得的两幅异常损失图采用取小运算得到最终的异常变化强度图,以便在减小输入图像间背景光谱差异的同时突出异常变化。在高光谱异常变化检测基准数据集上的实验结果表明,与10种相关方法相比,DSCAE展现了更优的检测性能。
